Well, I've looked in the source code for Hercules/390 again. It appears that
the B220 (SERVC) instruction, which is undocumented anywhere that I can
find, returns the "Service Call Control Block". One of the fields in the
SCCB, mapped by the IHASCCB macro in z/OS, is a byte field labelled SCCBRCC.
If bit 1 is "on" (value of .1... since S/390 labels the high order bit as
bit 0), then this is an IFL. Apparently z/OS, et al. check this bit early in
the IPL or NIP processing and, if on, load a wait state PSW.

Note - this does not necessarily mean that there are no other differences. I
would not attempt to "fake out" these licensed products by somehow messing
with this bit.

--
John McKown
Senior Technical Specialist
UICI Insurance Center
Applications & Solutions Team
+1.817.255.3225


> -----Original Message-----
> From: Bill Stermer [mailto:[EMAIL PROTECTED]
> Sent: Thursday, March 06, 2003 1:31 PM
> To: [EMAIL PROTECTED]
> Subject: Re: IFL speed.
>
>
> Hi Loren,
>
> From the meetings we had with IBM the IFL speed will remain
> at 192 mips. The basic difference between a CP engine and an
> IFL engine is that the microcode for the IFL engine is
> modified to only except Linux workloads not "legacy" stuff.
> If there is more magic at the hardware level it was not
> revealed in our meetings...
>
> Regards,
>
> Bill Stermer
> EDS - City of Anaheim

Reply via email to