When programatically taking an SVC dump, I incorrectly assumed that when
I did a ST REGS that I would see my registers at the time the SVC 51 was
issued, but they're all zeroes.
So I thought that I would find the SVRB and get the registers that way. So
I grab the TCB address from SYSTRACE, c
] On Behalf Of
Adam Johanson
Sent: 01 October 2009 14:13
To: IBM-MAIN@bama.ua.edu
Subject: SDUMPX and Regs at Time of Dump
When programatically taking an SVC dump, I incorrectly assumed that when I did
a ST REGS that I would see my registers at the time the SVC 51 was issued, but
they'r
Yeah and unfortunately, I'm having trouble with it. I'm getting a x'3908'
in R15 which means that the length of my parmlist is invalid. So maybe y'all
could help me there:
ocbitmask,incgpr
sdumpx hdrad=(2),
idad=(3),
type=failrc,
pswregs=regslist,
List [mailto:ibm-m...@bama.ua.edu] On Behalf Of
Adam Johanson
Sent: 01 October 2009 14:36
To: IBM-MAIN@bama.ua.edu
Subject: Re: SDUMPX and Regs at Time of Dump
Yeah and unfortunately, I'm having trouble with it. I'm getting a x'3908'
in R15 which means that the length of my parmlist
ST REGS is intended to show you the regs at time of error. It's pretty
unusual to want to display the regs at the time of the SVC Dump.
If it is showing 0's then in all likelihood the regs at time of error are
not available. Perhaps you are taking a dump out of "mainline" rather than
out of a reco
Rob Scott wrote:
I believe that the construction of the PSWREGS plist is cumulative (data
that you indicate via bit settings is not supplied is omitted rather than
being set to hex-zeros) - so if you just want the PSW+GPS then you need to
adjust the length to be 76 bytes and NOT include the control
: IBM-MAIN@bama.ua.edu
Subject: Re: SDUMPX and Regs at Time of Dump
Rob Scott wrote:
I believe that the construction of the PSWREGS plist is cumulative (data that
you indicate via bit settings is not supplied is omitted rather than being set
to hex-zeros) - so if you just want the PSW+GPS then y
On Fri, 2009-10-02 at 08:50 -0400, Rob Scott wrote:
> Thanks for the correction - looks like I missed a trick when reading the
> manual.
And had he access to the source I suspect that wouldn't have happened.
Mind you, a few of the rest of us might appreciate access to the source
(again) too ...
Peter,
Thanks for that explanation on PSWREGS. That explains my "PSWREGS
Woes" that I was having yesterday. I misunderstood the following from the
Friendly Manual:
"If a mask bit is off and any lower-order mask bit is on, the corresponding
storage area must be included in the parameter list
In , on 10/01/2009
at 08:13 AM, Adam Johanson said:
> So I thought that I would find the SVRB and get the registers that
>way. So I grab the TCB address from SYSTRACE, chain back the RBs and
>none of them have a WLIC for x'--020033' like I would expect.
What's in the PRB for the program t
> What's in the PRB for the program that issued the SDUMP?
SYSTRACE entry:
00 0045 007FF050 SVCR33
TCB:
007FF050 007FDBC0 007B5FE8 | .".{.#¬Y |
RBP field in TCB (4 bytes before this is shown since it's the WLIC):
007FDBBC
And then as an aside after I chased all those RBs I realized that I could
simply do a SUMM FORMAT and do a FIND for all the WLICs. I'm ready for the
weekend...
--
For IBM-MAIN subscribe / signoff / archive access instructions,
s
Adam Johanson quoted
"If a mask bit is off and any lower-order mask bit is on, the corresponding
storage area must be included in the parameter list."
That is from the z/OS R9 book. The z/OS R10 books are better, but still not
as clear as it could be:
"If a mask bit is set, SVC dump expects that
13 matches
Mail list logo
