On 2/20/2018 2:50 PM, Tony Thigpen wrote:
I am deeply saddened to report that John passed away this morning (Tuesday, 20
February 2018) following a protracted battle with cancer.
I am also deeply saddened. John tended to speak quietly and was a kind
and thoughtful person. Yet his work with
On 9/13/2017 12:26 PM, Charles Mills wrote:
Question: is there any way for the 'examining' code, possibly running on a
different CPU, to be assured of seeing a consistent (either 'before' or
'after') POINTER and ADDRESS contents? Possibly with serialization at some
point?
In general, no.
The
On 9/12/2017 9:35 AM, Charles Mills wrote:
Disabling for interruptions is not sufficient in a multi-processor world, right?
Disabling for interrupts guarantees, even in a multi-processing world,
that the executing unit of work will not lose control from the
perspective of the logical CP that
On 9/10/2017 11:39 AM, Charles Mills wrote:
...
Is that saying that there is a chance that another CPU might observe a
condition in which the store of the first operand had occurred, but the
store of the second operand had not? My understanding from earlier
paragraphs was that CSST happened "all
On 7/7/2017 7:44 AM, Binyamin Dissen wrote:
How big is the performance cost of running ARMODE?
Should there be a section of code that uses only ALET 0, is it best to go into
primary mode for that section?
Nothing, so long as AR translation is not required. And an ALET of 0
(or 1) would
On 7/7/2017 7:44 AM, Binyamin Dissen wrote:
What is the performance cost of using key 9 storage when the PSW is not in key
9?
Nothing (zip, nil, nada).
Regards,
Greg
--
For IBM-MAIN subscribe / signoff / archive access
On 6/23/2017 11:37 AM, contactmura...@gmail.com wrote:
Hi, Thanks for the response. But while testing it is observed that if we
execute FRR with a debugger then it doesn't work as expected. Mostly because
the debugger places its own FRR in the FRR stack.
That sounds like a concern to take up
On 6/22/2017 4:17 AM, contactmura...@gmail.com wrote:
I am new to SRBs & FRRs. I have a question regarding FRR Parameter address
discussed over here.
When said "If the SRB abends, its FRR receives the address of the FRR parameter area
in R2 (or SDWAPARM)."
Why is it (or SDWAPARM)? That means
On 6/19/2017 2:02 AM, contactmura...@gmail.com wrote:> I was expecting
'SRB Abended, Inside FRR now' to be displayed before displaying
'Returned from SRB. In TCB mode now'. Because as per my initial
understanding of SYNCH=YES, TCB will be suspended until SRB completes.
But as soon as the abend
On 6/19/2017 11:51 AM, Donald Likens wrote:
Does anyone see any problems with the following code... Updating free space
pointer.
To fully understand the possibilities for error you would need to show
both sides of the processing, allocation and freeing. In cases like
this there is almost
On 6/19/2017 8:00 AM, Todd Arnold wrote:
- If you need "secure keys" - keys that are protected by hardware that
cannot be subverted, even by the highest-technology methods - then
use CEX. (but if you need a lower level of security, consider CPACF
Protected Key mode.)
I would note
On 6/13/2017 2:55 AM, David Cole wrote:
PRs do automagically cancel all ESTAEs created since the matching BAKR.
There is a bit (in the BAKR's stack entry I think. I forget which and
where.) that, when set on, causes an interrupt to occur when the PR is
issued. This is how z/OS gains control so
On 6/9/2017 11:47 AM, Richards, Robert B. wrote:
I seem to remember that V=R was used with MVS/XA running under VM/XA SF and you
could bounce VM/XA SF and the guest would stay up.
That was a *VM* V=R user, not a MVS V=R address space.
Regards,
Greg
On 6/8/2017 10:44 AM, John McKown wrote:
This still seems to be supported in z/OS 2.2. Does anyone need to run a
program V=R in today's world? I'm just curious because this support seems
to be a "waste" of protect keys 10 through 15. Of course, if those keys
were "freed up", what could they be
On 6/2/2017 10:59 AM, Joseph Reichman wrote:
To clarify storage if not released is around for the life of the TCB I am
assuming TCB ASCBXTCB is around for the life span of the address space
The answer is no, so step carefully.
ASCBXTCB is *not* a constant value for the life of the address
On 5/25/2017 10:38 PM, Donald Likens wrote:
SYNCH-YES does not say what happens when the SRB is suspended.
I am hoping that the work unit will resume when the SRB is suspended. Now that
I think about it, I do not think I should do this unless IBM says it will work
(or you all say it will work
On 5/15/2017 11:27 AM, Paul Gilmartin wrote:
What does the TSO TMP use to accomplish this?
Extreme care ;-)
It has been a while, but my memory is that the TMP stops all of the
tasks above (or is that below?) it in the task tree and then passes the
request to a special jobstep task (with
Be aware that what you are attempting to do is dangerous and has the
potential to create system integrity exposures that would allow a
problem state program to cause a system failure. I am not saying that
it can not be done safely, because it can be. But to do it safely
without creating a
On 4/22/2017 10:24 AM, esst...@juno.com wrote:
Can I issue a SETRP DUMP=YES without other parameters.
.
Meaning I don't want to specify a retry routine nor do I want to specify a
return code.
I simply want to issue SETRP DUMP=YES and return to the next sequential
instruction following
the
On 4/18/2017 2:25 AM, Vernooij, Kees - KLM , ITOPT1 wrote:
As I said, I remember reading this a long time ago, I don't know the details
anymore, whether the source was reliable and whether it is still working this
way. Only a DB2 internal expert should be able to tell.
...
On 12 April 2017
On 4/11/2017 3:26 PM, Paul Gilmartin wrote:
My understanding, ancient, probably outdated, and certainly naive is that
there is little communication between GETMAIN/FREEMAIN and the paging
subsystem. If a program touches a page that was never GETMAINed no error
occurs; simply a page slot is
On 4/11/2017 1:46 PM, Jesse 1 Robinson wrote:
Part of the problem, I learned some time back at SHARE, is that there is no
mechanism to 'reclaim' page slots that no longer need to remain on disk. Once
storage gets paged out, it sits there like a sandbag until the owning task is
stopped.
On 4/11/2017 2:24 PM, Blaicher, Christopher Y. wrote:
It has been a while since I worked on DB2, but it is sounding like your buffer
pools are too big.
With large memory systems everyone should have all of the production
(and, ideally, test systems too) buffer pools define with PGFIX(YES)
On 3/27/2017 3:59 PM, Paul Gilmartin wrote:
It has been discussed here for a while. You could disable interrupts, branch to
code above the bar, and branch back later. (I suppose the Old PSW was
unconditionally scrunched.) More recently, interrupts above the bar are
tolerated, but no system
On 3/16/2017 2:30 AM, kalda0...@gmail.com wrote:
> When I try to format the trace in IPCS the following error is received:
>
> BLS21082I DSNAME('SYS1.TRACE')I/O error
> BLS21081I 0AB1,D,SYS00010,READ,OUT OF EXTENT,TTR UNKNOWN,BSAM
> AHL10004I Input is not a GTF trace dataset.
> AHL10009I No
On 3/13/2017 6:00 AM, Barbara Nitz wrote:
According to the books, IEF743I means "System action: The job and address space
end." Unfortunately that does not seem to be the case. There is no indication that I
can find in hardcopy log that the address space (Initiator) got restarted, and the dump
On 3/6/2017 3:15 PM, Jim wrote:
I've used TBEGIN and vaguely remember reading up on TBEGINC but never ever
tried it .. I'll have to hit the
foils again ..
My recommendation is that whenever possible, use a constrained
transaction (aka, TBEGINC). It is guaranteed to complete successfully,
On 2/22/2017 2:19 PM, John McKown wrote:
​But I agree that // ELSIF or // ELSEIF would be superior. DEFAULT /
OTHERWISE is simply after the last ELSE. Perhaps instead of // IF, IBM
should have have // CASE?
Ahhh, but will your company *pay* IBM more $$ for z/OS if this
functionality were
On 2/22/2017 12:57 PM, esst...@juno.com wrote:
That other programs can read but not update.
.
Lets say I use KEY 4 as an example.
If I issue a STORAGE OBTAIN with key 4 the system should use
a default subpool. From an Integrity perspective is it acceptable
to use the default subpool or would it
On 2/22/2017 10:15 AM, Joseph Reichman wrote:
I know if you specify GSPV or SHPSV on the attach tasks can Share ? A subpool
If this is not specified on the attach would a subtask get a S0C4 for
referencing storage
Obtained by the originating task
No.
A subpool is a logical grouping of
On 2/22/2017 7:17 AM, Joseph Reichman wrote:
I am trying to reference storage that I ( my TCB didn't necessarily obtain )
Sometime I think if this storage has been paged out I would get a S0C4 pic 11
If VSMLIST determines it has been getmain'ed
Maybe I can PGSER FIX back in
I just got 2.2
On 2/6/2017 11:46 AM, Peter Hunkeler wrote:
- Are the "DB2 latches" implemented as GRS latches?
- Are row, page, tables space locks actually DB2 latches? In other words, would
WLM be able to recognize a low priority job holding some DB2 locks is causeing
delay so it can promote it?
- What have
On 2/2/2017 8:20 AM, Blaicher, Christopher Y. wrote:
I think this guy has something he wants to sell because while PARS was
developed on a 7074, in the early 1970's it was evolved by IBM into ACP and TPF.
The 360/65 and some other 360 and 370 machines had the capability to run a 7074
On 1/19/2017 9:24 AM, Greg Dyck wrote:
Many, many, moons ago type 2 SVCs were managed differently from type 4
SVCs. It's a fuzzy memory, but I believe they had to be 2K or less in
size and were loaded from SVCLIB transiently into a common 2K area. That
lead to restrictions for them
On 1/19/2017 7:55 AM, Joseph Reichman wrote:
I used to work for Vendors I now work in applications Assembler
I write internals code to keep those skills sharp
If you look at the doc for in the Authorized Assembler guide I believe chap 23
It says type 2 SVC can not suspend their caller
I am
On 1/11/2017 3:30 PM, Frank Swarbrick wrote:
Is there a downside to always defining VSAM files with a MAXLRECL of 32761,
which seems to be the largest value for this parm for an UNSPANNED dataset?
I believe it will force your CI size to be 32K which is not something I
would want to do unless
On 1/10/2017 11:30 AM, Charles Mills wrote:
The single TIOT is a limit to "virtualization."
Certainly the function could be useful. While conceptually it seems
easy to do, where the rubber meets the road in allocation, access
methods, SMF accounting, and more, it would be a very
On 12/22/2016 1:47 AM, Peter Hunkeler wrote:
When I get control, R13 points to a save area, which my code will use, but I do
not yet have storage for another save area before calling IEANTRT, so I cannot
change R13.
The Name Token Retrieve service uses BAKR to save and restore the
caller's
On 12/21/2016 2:59 PM, Charles Mills wrote:
When does ECTG come along? I've got high-level notes on the various z architecture levels
but the "extract-CPU-time facility" is not in my notes.
For those who still remember S360 architecture ;-), it is a recent
addition. z9-109, circa September
On 12/21/2016 10:46 AM, Paul Gilmartin wrote:
How does ECTG do that? Is there information in control registers supporting it?
Does it work likewise on systems other than z/OS, such as Linux? VM guests?
The high level answer is that ECTG atomically fetches 8 bytes of storage
from one
On 12/20/2016 5:58 PM, Bill Woodger wrote:
If there's some way to make it more "accurate" for what I want, I'd be pleased.
Use the assembler macro-
TIMEUSED STORADR=xxx,ECT=YES
This will call a *short* system routine (no PC, no SVC) which will use
the ECTG instruction to calculate the
Is there some formula to calculate a MIPS usage from the next kind of
data-source and volumes estimation:
Keep in mind that MIPS stands for "Meaningless Indicator of Processor
Speed"... long used, usually misunderstood, and often abused. On a
current processor I can write one program that
On 10/19/2016 7:51 AM, Peter Hunkeler wrote:
But the net seems to be, my assumption is wrong. If an area >4k is PGFIXed,
then I need to find the real address of every page and cannot assume the real
storage is contiguous.
This is a correct statement. PGFIXing storage guarantees the page is
You *can* *not* safely intermix the use of CDSG and PLO without using
special protocols. This is clearly documented in the POPs.
You *can* safely intermix CDSG and transactional execution.
Regards,
Greg
--
For IBM-MAIN
On 10/7/2016 1:38 PM, Jim Mulder wrote:
Since memterm does not access the storage of the address
being terminated, there is no connection between IEFUSI and memterm.
There is no requirement for any available storage in the address
space being memtermed. Task termination, yes.
Memory
On 9/29/2016 4:18 PM, Jesse 1 Robinson wrote:
Ah, that's a point I'm still fuzzy on. When you say 'Control Areas first used
after CA-Reclaim is turned on', does that mean that reclaim will not affect an
old cluster until it has been (re)created? That makes a difference in how to
roll it out.
On 9/29/2016 10:17 AM, Peter Hunkeler wrote:
So there should be enough spare capacity for the system (LPAR) to to use. But
it does not. The 7 vertical low CPs are mostly parked or unparked for only a
few percent. I wonder why MVS is not using more CPs and more Capacity for that
job.
You
On 9/28/2016 11:29 AM, Peter Hunkeler wrote:
Job in question is DB2 reorg utlility job which runs some 30 subtasks in
parallel runs for 2 hours. RMF III (60s intervals) shows that the job mostly
has a good workflow (80%+), is heavily using CP (often 80%-90%) and at the same
time is heavily
On 9/13/2016 11:55 AM, Jake Anderson wrote:
Can some one please explain the Catalog internals with respect to the VVDS
and VTOC.
At the 10,000 ft level...
The catalog is a VSAM KSDS dataset that (mostly) points to the volume
containing the dataset.
The VTOC contains where on the volume, by
On 8/6/2016 9:49 AM, Paul Gilmartin wrote:
I don't believe there's a timing hazard; DYNALLOC will go to the end of the name
space and wrap before it reuses the DDNAME you got.
Dynamic allocation *will* *not* reuse the DDNAME. After generating a
name it checks to see if it is already in use,
On 8/3/2016 1:10 PM, Charles Mills wrote:
I am trying to shoot a problem. The code in question makes reference to a
TCB shortly after issuing a DETACH for that TCB. Am I correct in my analysis
that that is a S0C4 candidate? That once you issue a DETACH, the TCB is no
longer necessarily valid?
On 7/28/2016 1:00 PM, Joe Reichman wrote:
I have a FRR covering a SRB it not a parameter on the SCHEDULE or IEAMSCHD
but rather invoked inside the SRB. The documentation says to delete it just
code SETFRR D,WORKREGS However I am guessing it has to be in the context of
the SRB or in the FRR on
On 6/21/2016 1:11 PM, Lizette Koehler wrote:
After the discussion, I went ahead and put in a RFE for IEF240I TIOT Exceeded
message.
http://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe_ID=90368
Knowing the amount probably won't help you resolve a concern, given the
64K hard limit on
On 6/18/2016 11:03 AM, Jake Anderson wrote:
Is there a way to know if the DD chain has approached the end or it has
approached the end of Chain. Is there a way to tweak the JCL to overcome
TIOT exceeded values ?
The TIOT is a fixed sized array of entries, not a chain. The size is
defined in
On 6/13/2016 11:18 AM, Charles Mills wrote:
Does that mean that for standard "old-fashioned" AMODE 31 72-byte savearea
linkage I am obligated to save the high word of R1 before issuing LINK in a
called program? If so, this would seem to be a compatibility issue for older
code that uses LINK. Or
On 6/8/2016 7:07 AM, michelbutz wrote:
The XMEM post is being done from a FRR resulting from an abend in a SRB the SRB
is the primary address space
Putting an ECB in common storage is allowed, but great care must be
taken when doing so.
As has already been said, if issuing the POST from
On 6/7/2016 8:41 AM, michelbutz wrote:
Can someone tell me what the advantage of this
Is then using IEAVPSE2 is it performance
I think they both set up a SSRB
SUSPEND/RESUME w/TOKEN *only* works for SRBs. Pause/Release works for
both tasks and SRBs. Performance wise, SUSPEND/RESUME w/Token
On 6/6/2016 6:54 AM, Steve Smith wrote:
> I'm wondering what an SRB SUSPEND exit is really good for. The
Services Guide and Reference state that it is required, and that it is
supposed to save the SUSPEND token, and decide whether to allow the
SUSPEND. I can't see why the SRB mainline
On 6/4/2016 6:34 AM, Ed Jaffe wrote:
So, the local lock is obtained in only a subset of Pause/Release scenarios?
Ed,
The local lock is *never* obtained for single PET Pause/Release
processing.
For multiple PET Pause/Release it is obtained to serialize creation of
an address space related
On 6/3/2016 2:24 AM, michelbutz wrote:
Does it have to be a different value on every call
If I understand your question, the answer is no.
You *might* use a different value to indicate the PET was released for a
specific reason, such as 0 for process work, 2 for error occurred, and 3
for
On 6/2/2016 9:43 PM, michelbutz wrote:
So it is the same parm as the one in the pause
Yes.
. IEAVPSE/IEAVPSE2/IEA4PSE/IEA4PSE2 release_code
. IEAVXFR/IEAVXFR2/IEA4XFR/IEA4XFR2 current_du_release_code
will be set to the value specified by one of the following-
.
On 6/2/2016 7:08 PM, michealbutz wrote:
I have a question about the 3rd paramter of IEAVRLS2 target_du_release_code
it is specfied as 4 bytes. Is this the same as the 4th paramter of the Pause
service release code this is specfied as being 3 bytes
Sigh... I checked the code and the description
On 5/29/2016 10:24 PM, Peter Hunkeler wrote:
Thanks, Greg, for the pointer to the presentation. I did not read anything
which would contradict my understanding. I conclude below text, which I cited
previously, needs some rework. I' ll submit an RCF.
> Back to the OP's topic. I read in
On 5/26/2016 4:24 PM, Charles Mills wrote:
So this is inherent in the guts of MVS -- not something in PROGxx or
somewhere like that?
Yes, it is inherent in the guts of MVS. There is no external knob to
change the initial value that a jobstep task starts at. But there are
multiple ways that
On 5/26/2016 11:19 PM, Peter Hunkeler wrote:
This makes my wonder. My understanding, which maybe is not entirely correct, is
that the dispatcher only looks at the work unit queue (WUQ) for the type of
processor, when a processor needs new work. It picks the first work element
block (WEB) form
On 5/26/2016 2:51 PM, Charles Mills wrote:
1. Where is the original dispatching and limit priority for the jobstep task
set? How do I determine (not with an API at runtime, but in advance with a
command or looking at a parm member) what it is presumably going to be?
Sigh... the initial
On 5/26/2016 10:29 AM, Jerry Callen wrote:
I wrote a simple test program to compare the performance of WAIT/POST and pause elements.
The program has two tasks and simply ping-pongs back and forth between them (no
overlapped execution). Each task has a synchronization gadget, either an ECB or a
On 5/25/2016 11:16 AM, Peter Hunkeler wrote:
Yet still the count field is printed on a line by its own. This is why I had
expected the key field to printed separately from the data field as well.
Someone made a decision 30 years ago on how DFDSS PRINT should format
the record information,
On 5/19/2016 9:54 AM, Phil Smith III wrote:
...
The surprise here was that since R14 had something in the top half of the
grande register, Bad Things resulted (S0C4) on the line with comment "ADDR
SYST LINKAGE TABLE". This seemed.unintuitive. Are we the only ones who were
(or have been)
On 5/12/2016 10:17 AM, Mark Zelden wrote:
On Thu, 12 May 2016 07:30:03 -0400, Richards, Robert B.
wrote:
I do have a follow-up question Is anyone aware of a downside to over
specifying reserved CPs?
LCCA/PCCA storage. But the default CBLOC has been VIRTUAL31
On 5/4/2016 10:41 AM, Carlos Bodra wrote:
SOURCE AND TARGET DEVICE CAPACITIES DO NOT MATCH. CYLINDER CAPACITY OF
SOURCE VOLUME 2721, TARGET VOLUME 0D0B.
This tells me that source volume has 2721 cylinders and target volume
has 3339 cylinders (0d0b).
No, the source volume has 10017 cylinders
On 5/3/2016 4:30 AM, Peter Hunkeler wrote:
I'm analyzing a job using CA's MA-Tune (similar to Strobe). Question is if
there is that could be optimized. Looking at the MA-Tune reports which are
based on 100 samples per second for 1 minute, I see that the job is seen in
IEAVEPS1 for some 20%.
On 5/1/2016 3:45 PM, michelbutz wrote:
Hi
If a nonpreamtable scope=global SRB does a pause
And another SRB is scheduled does it get dispatched ?
Being non-preemptable means that z/OS won't take control away from the
unit of work without cause. Cause being something like a page fault
On 5/2/2016 12:54 PM, Blaicher, Christopher Y. wrote:
What parameters were on the IEAMSCHD macros for the two SRB's? If the
LLOCK=YES parm is set on both, then there is your problem.
A Pause is not allowed while holding *any* class of lock. The Pause
request will fail.
Greg
Tony,
If you can get into supervisor state, use the STSI instruction.
Under z/OS the CSRSI service uses STSI to provide the data that it
returns to problem program state requestors. I don't know if VSE
provides any equivalent service.
Greg
On 4/25/2016 10:55 AM, Tony Thigpen wrote:
Is
On 4/20/2016 7:25 PM, Steve Beaver wrote:
DISASM is a nice little "gimme" but I don't think it's been updated in
years. You might find yourself
Doing a lot of coding of DISASM to get it functional for 32 and 64 bit code
I believe you are thinking about a different "DISASM" than the support
How to determine the length of the block for a BSAM/BPAM RECFM=FB
dataset is fully described in the z/OS DFSMS Using Data Sets book,
Processing Sequential Data Sets, Determining the Length of a Block.
Greg
What big and new (30 years old or less) companies use mainframe? Google?
Facebook? Netflix?
Maybe not directly, but they are certainly dependent on banks, insurance
companies, the SEC, and others to accurately, quickly, securely, and
with integrity and safety, maintain information about
Rule 3 says there are events, some explicit (BCR 15,0) and some implicit
(interrupts), that force conceptual completion to occur, over and above
those required by rule number 1.
These rules are for storage consistency, especially between multiple
processors.
Registers are slightly different
Operand fetches and stores must appear to occur in proper order.
So is it saying that the machine obeys Rule 1? I will bet
dollars to donuts that that is not the case, that the machine
performs read operations before logically preceding write
operations, and that a program can be run to
As to the question, what can you do with the information CSVFETCH makes
available: can I write an SMF record from CSVFETCH?
You should be able to use SMFEWTM BRANCH=YES can write a SMF record from
the exit.
Greg
--
For
On 1/24/2015 9:33 AM, J O Skip Robinson wrote:
Yikes. I don't fully understand the impact of OA44240. The APAR text does not
mention UCB lookup. We have several RYO and an indeterminate number of ISV
programs that utilize the 'traditional' UCB lookup process. Will everything
fall apart when
From what we can see the address that ends up in the VCON is the
displacement from the beginning of the module (the value presumably
placed their by the binder) plus twice the load point. Has anyone seen
this happen before? Is there an undocumented requirement to single
thread multiple LOADs
On 1/14/2015 12:28 PM, Mike Schwab wrote:
(So it looks like any individual task that is CPU bound and limited to
one core will take twice as long.)
What you state is always true for temporal multithreading, which
timeslices the execution of CP x with CP y on a single core.
But for
On 1/12/2015 7:52 AM, Donald Likens wrote:
Can anyone explain why the following code did not work in AMODE 64 but works in
AMODE 31?
Yes... bit 32 of R15 on entry to your STIMER exit routine is on. In
AMODE 64 the LA of the ECB address propagates it into R1. Thus R1
indicates it is a
Make sure the storage for the DCBE has the same storage key as the DCB,
contains the constant 'DCBE' in the first 4 bytes, and has a minimum
length value of 56 in the halfword that follows the constant 'DCBE'.
Greg
--
For
On 1/6/2015 8:47 AM, Binyamin Dissen wrote:
SDUMPX DCB=OSVCDDCB, +
ECB=(DUMPECB,WRITE),+
SUMLIST64=SUMLIST64,+
LIST64=(R7),
87 matches
Mail list logo