Notes For V3.9

The makefile now works for Linux and most Unix's. However, for Solaris
and MacOS, you must first export the OSTYPE environment variable:

> export OSTYPE
> make

Otherwise, you will get build errors.

1. New Features

1.1 3.9-0

1.1.1 SCP and libraries

	- added *nix READLINE support (Mark Pizzolato)
	- added "SHOW SHOW" and "SHOW <dev> SHOW" commands (Mark Pizzolato)
	- added support for BREAK key on Windows (Mark Pizzolato)

1.1.2 PDP-8

	- floating point processor is now enabled

2. Bugs Fixed

Please see the revision history on http://simh.trailing-edge.com or
in the source module sim_rev.h.

3. Status Report

This is the last release of SimH for which I will be sole editor. After this
release, the source is moving to a public repository:

under the general editorship of Dave Hittner and Mark Pizzolato. The status
of the individual simulators is as follows:

3.1 PDP-1

Stable and working; runs available software.

3.2 PDP-4/7/9/15

Stable and working; runs available software.

3.3 PDP-8

Stable and working; runs available software.

3.4 PDP-10 [KS-10 only]

Stable and working; runs available software.

3.5 PDP-11

Stable and working; runs available system software. The emulation of individual
models has numerous errors of detail, which prevents many diagnostics from
running correctly.

3.6 VAX-11/780

Stable and working; runs available software.

3.7 MicroVAX 3900 (VAX)

Stable and working; runs available software. Thanks to the kind generosity of
Camiel Vanderhoeven, this simulator has been verified with AXE, the VAX
architectural exerciser.

3.8 Nova

Stable and working; runs available software.

3.9 Eclipse

Stable and working, but not really supported. There is no Eclipse-specific
software available under a hobbyist license.

3.10 Interdata 16b

Stable and working, but no software for it has been found, other than
diagnostics.

3.11 Interdata 32b

Stable and working; runs 32b UNIX and diagnostics.

3.12 IBM 1401

Stable and working; runs available software.

3.13 IBM 1620

Hand debug only.  No software for it has been found or tested.

3.14 IBM 7094

Stable and working as a stock system; runs IBSYS. The CTSS extensions
have not been debugged.

3.15 IBM S/3

Stable and working, but not really supported. Runs available software.

3.16 IBM 1130

Stable and working; runs available software.  Supported and edited by
Brian Knittel.

3.17 HP 2100/1000

Stable and working; runs available software. Supported and edited by
Dave Bryan.

3.18 Honeywell 316/516

Stable and working; runs available software.

3.19 GRI-909/99

Hand debug only.  No software for it has been found or tested.

3.20 SDS-940

Hand debug only, and a few diagnostics.

3.21 LGP-30

Unfinished; hand debug only. Does not run available software, probably
due to my misunderstanding of the LGP-30 operational procedures.

3.22 Altair (original 8080 version)

Stable and working, but not really supported. Runs available software.

3.23 AltairZ80 (Z80 version)

Stable and working; runs available software. Supported and edited by
Peter Schorn.

3.24 SWTP 6800

Stable and working; runs available software. Supported and edited by
Bill Beech

3.25 Sigma 32b

Incomplete; more work is needed on the peripherals for accuracy.

3.26 Alpha

Incomplete; essentially just an EV-5 (21164) chip emulator.

4. Suggestions for Future Work

4.1 General Structure

	- Multi-threading, to allow true concurrency between SCP and the simulator
	- Graphics device support, particularly for the PDP-1 and PDP-11

4.2 Current Simulators

	- PDP-1 graphics, to run Space War
	- PDP-11 GT40 graphics, to run Lunar Lander
	- PDP-15 MUMPS-15
	- Interdata native OS debug, both 16b and 32b
	- SDS 940 timesharing operating system debug
	- IBM 7094 CTSS feature debug and operating system debug
	- IBM 1620 debug and software
	- GRI-909 software
	- Sigma 32b completion and debug
	- LGP-30 debug

4.3 Possible Future Simulators

	- Data General MV8000 (if a hobbyist license can be obtained for AOS)
	- Alpha simulator
	- HP 3000 (16b) simulator with MPE

VAX/vax_octa.c

@@ -1,6 +1,6 @@
 /* vax_octa.c - VAX octaword and h_floating instructions
 
-   Copyright (c) 2004-2008, Robert M Supnik
+   Copyright (c) 2004-2011, Robert M Supnik
 
    Permission is hereby granted, free of charge, to any person obtaining a
    copy of this software and associated documentation files (the "Software"),
@@ -25,6 +25,9 @@
 
    This module simulates the VAX h_floating instruction set.
 
+   15-Sep-11    RMS     Fixed integer overflow bug in EMODH
+                        Fixed POLYH normalizing before add mask bug
+                        (both from Camiel Vanderhoeven)
    28-May-08    RMS     Inlined physical memory routines
    10-May-06    RMS     Fixed bug in reported VA on faulting cross-page write
    03-May-06    RMS     Fixed MNEGH to test negated sign, clear C
@@ -93,7 +96,7 @@ void h_write_w (int32 spec, int32 va, int32 val, int32 acc);
 void h_write_l (int32 spec, int32 va, int32 val, int32 acc);
 void h_write_q (int32 spec, int32 va, int32 vl, int32 vh, int32 acc);
 void h_write_o (int32 spec, int32 va, int32 *val, int32 acc);
-void vax_hadd (UFPH *a, UFPH *b);
+void vax_hadd (UFPH *a, UFPH *b, uint32 mlo);
 void vax_hmul (UFPH *a, UFPH *b, uint32 mlo);
 void vax_hmod (UFPH *a, int32 *intgr, int32 *flg);
 void vax_hdiv (UFPH *a, UFPH *b);
@@ -581,7 +584,7 @@ h_unpackh (&opnd[0], &a);                               /* unpack s1, s2 */
 h_unpackh (&opnd[4], &b);
 if (sub)                                                /* sub? -s1 */
     a.sign = a.sign ^ FPSIGN;
-vax_hadd (&a, &b);                                      /* do add */
+vax_hadd (&a, &b, 0);                                   /* do add */
 return h_rpackh (&a, hflt);                             /* round and pack */
 }
 
@@ -643,7 +646,7 @@ for (i = 0; i < deg; i++) {                             /* loop */
     wd[3] = Read (ptr + 12, L_LONG, RD);
     ptr = ptr + 16;
     h_unpackh (wd, &c);                                 /* unpack Cnext */
-    vax_hadd (&r, &c);                                  /* r = r + Cnext */
+    vax_hadd (&r, &c, 1);                               /* r = r + Cnext */
     h_rpackh (&r, res);                                 /* round and pack */
     }
 R[0] = res[0];                                          /* result */
@@ -678,7 +681,7 @@ return h_rpackh (&a, hflt);                             /* round and pack frac *
 
 /* Floating add */
 
-void vax_hadd (UFPH *a, UFPH *b)
+void vax_hadd (UFPH *a, UFPH *b, uint32 mlo)
 {
 int32 ediff;
 UFPH t;
@@ -703,6 +706,7 @@ if (a->sign ^ b->sign) {                                /* eff sub? */
     if (ediff)                                          /* denormalize */
         qp_rsh_s (&b->frac, ediff, 1);
     qp_add (&a->frac, &b->frac);                        /* "add" frac */
+    a->frac.f0 = a->frac.f0 & ~mlo;                     /* mask before norm */
     h_normh (a);                                        /* normalize */
     }
 else {
@@ -713,6 +717,7 @@ else {
         a->frac.f3 = a->frac.f3 | UH_NM_H;              /* add norm bit */
         a->exp = a->exp + 1;                            /* incr exp */
         }
+    a->frac.f0 = a->frac.f0 & ~mlo;                      /* mask */
     }
 return;
 }
@@ -778,7 +783,14 @@ else if (a->exp <= (H_BIAS + 128)) {                    /* in range? */
     a->exp = H_BIAS;
     }
 else {
-    *intgr = 0;                                         /* out of range */
+    if (a->exp < (H_BIAS + 160)) {                      /* left shift needed? */
+        ifr = a->frac;
+        qp_lsh (&ifr, a->exp - H_BIAS - 128);
+        *intgr = ifr.f0;
+        }
+    else *intgr = 0;                                    /* out of range */
+    if (a->sign)
+        *intgr = -*intgr;
     a->frac.f0 = a->frac.f1 = 0;                        /* result 0 */
     a->frac.f2 = a->frac.f3 = 0;
     a->sign = a->exp = 0;