[Simh] RRD40 emulation problem

[Mike Duffy]

Not too long ago, I modified SIMH to directly access a physical host CD 
drive, in order to eliminate the need to make a disk image before using a 
CD.  (Migration Specialties asked me to do this, with the intention of 
donating it back to the SIMH community.  We've supplied a copy of the source 
to be evaluated by Mr. Supnik, but have not heard anything back yet.)

While I was in there, I noticed tables for each of the supported device 
types containing hardcoded values for blocks, cylinders, tracks, etc., and I 
didn't notice any provision for detecting different values on the fly.  The 
exception was RAUSER, which as far as I noticed, only changed one field 
(size), not the whole geometry.  Based on what I saw, it should not be too 
hard to modify the ATTACH code to detect the image size and change that same 
field when the device type is RRD40 (or my new type, PHYCD).

One potential problem could be that for some reason this change really does 
need to be done at SET time rather than at ATTACH time.  That might make my 
idea difficult or impossible.  I don't know of any such reason at this time.

Would you feel comfortable trying to make such a change?  I can get out my 
notes and try to provide a little information if you'd like.

(For my own selfish purposes, it would also be nice to know whether the 
Physical CD code would exhibit the same error as the base SIMH code if 
supplied with the disk in question.)

-Mike Duffy



----- Original Message ----- 
From: <del121 at cinci.rr.com>
To: <simh at trailing-edge.com>
Sent: Wednesday, April 08, 2009 11:11 AM
Subject: [Simh] RRD40 emulation problem


>I recently ripped a number of CDs, including a 'vms 5.5-2 remastered' cd.
>
> This particular cd image would not mount on simh-vax running vms 6.2, 
> giving an
> 'unsupported structure level' error. However, the actual cd mounts fine on 
> my
> uVAX 3100 M40 running vms 6.2
>
> The most obvious difference between this cd rip and the others was it's 
> size;
> the 5.5-2 cd rip was under 200MB, all my other CD rips were over 400MB.
>
> Operationally the most obvious difference between 'inserting' the cd on 
> the
> simhvax and my 3100 was the vms device attributes. On simh-vax VMS always
> reports 'Total Blocks 1331200' but on the 3100 it seems the hardware 
> senses the
> actual track size; differant CDs yield differant 'Total Blocks'.
>
> On a hunch, I changed my simh device type from RRD40 to RA60 (~200MB 
> removable
> disk pack) and attached the *same* 5.5-2 cd rip (readonly) to that. Voila! 
> Now
> VMS happily mounts it on simh-vax.
>
> Additional info - ANA/DISK of *all* ODS2 CD rips I tested on simh-vax 
> rrd40
> yields an error message about bitmap.sys being wrong, but NO such error 
> with
> real cds on mv3100.
>
> Putting this all together, I conclude there is a deficiency in the RRD40
> emulation. My guess is that simh needs to determine the size of the cd rip 
> when
> the image is attached and report actual
> total blocks when vms asks...
>
> FYI, the above is with reference to the Windows prebuilt-with-Ethernet 
> 3.8-1
> executable.
>
> Oh yeah, side note/questions -
>
> Does the rrd40 emulation throttle the transfer rate to 150KB/S (1X CD)? 
> Have not
> done timed tests but indirect evidence suggests that it does. This is not 
> a
> complaint, just curious if what I sensed was actually the case. (If so, 
> perhaps
> a 2X rrd42 could be added?)
>
> Further, if it *is* the case that the differant mass storage devices 
> simulate
> their associated device transfer characteritics, which simulated disk is 
> the
> fastest? RA7x, RA8x, RA9x ?
>
> Thanks...
>
> _______________________________________________
> Simh mailing list
> Simh at trailing-edge.com
> http://mailman.trailing-edge.com/mailman/listinfo/simh