I guess I conflated RTM and SVC Dump. It's not like I've ever seen either.
:) The real to virtual part is just a guess I made if ASID 4 was included
in the dump. If the IHATTE blocks are copied to the abending address space
before dumping, I don't see how to find them. The data area manual only
seems to be talking about how to find them in ASID 4.

In article <[email protected]> you wrote:
> > RTM copies the trace info from ASID 4 to the dump. If you want to look 
> at the
> > raw buffers, I think you'd need to find some way to include ASID 4 
> > in the dump 
> > and get to the buffers via control reg 12 and figure out how to convert 
> the
> > real address to a virtual. I now remember looking at IHATTE for some 
> reason
> > and that's the macro for mapping the trace entries. Most of this stuff 
> isn't
> > GUPI.

>   RTM does not copy the trace info from ASID 4 to the dump.  The SVC Dump,
> IEATDUMP, and SYSMDUMP programs use an internal system trace service to 
> copy data from the trace snapshot in the TRACE address space into their 
> own storage so that they can dump it.  That processing uses virtual 
> addresses.  No conversion from real to virtual is needed.  The low order
> part of the real address in CR12 (or a snapshot of it in a TBVT) is used
> to during formatting to determine the current position within the current 
> trace buffer for each CPU.  The high order part of the real address in 
> CR12
> is used only by the machine in order to determine where to put a trace 
> entry.

>    Jim Mulder 

-- 
Don Poitras - SAS Development  -  SAS Institute Inc. - SAS Campus Drive
[email protected]           (919) 531-5637                Cary, NC 27513

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to