Rivoreo Source Code Repositories
src.rivoreo.one / emulators / simh / 9c930092a4ff41cc5775e165cffcd42cee2b6c30 / . / H316 / h316_cpu.c
blob: 813fe077891c402546f29e34cc50eccdb6a4103f [file] [log] [blame] [raw]
/* h316_cpu.c: Honeywell 316/516 CPU simulator
Copyright (c) 1999-2015, 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"),
to deal in the Software without restriction, including without limitation
the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
ROBERT M SUPNIK BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Except as contained in this notice, the name of Robert M Supnik shall not be
used in advertising or otherwise to promote the sale, use or other dealings
in this Software without prior written authorization from Robert M Supnik.
cpu H316/H516 CPU
21-May-13 RLA Add IMP/TIP support
Move SMK/OTK instructions here (from CLK)
Make SET CPU DMA work as documented
Implement extended interrupts
Add "interrupt taken" flag to CPU HISTORY
Add "break on write" breakpoints
19-Nov-11 RMS Fixed XR behavior (Adrian Wise)
19-Nov-11 RMS Fixed bugs in double precision, normalization, SC (Adrian Wise)
10-Jan-10 RMS Fixed bugs in LDX, STX introduced in 3.8-1 (Theo Engel)
28-Apr-07 RMS Removed clock initialization
03-Apr-06 RMS Fixed bugs in LLL, LRL (Theo Engel)
22-Sep-05 RMS Fixed declarations (Sterling Garwood)
16-Aug-05 RMS Fixed C++ declaration and cast problems
15-Feb-05 RMS Added start button interrupt
01-Dec-04 RMS Fixed bug in DIV
06-Nov-04 RMS Added =n to SHOW HISTORY
04-Jan-04 RMS Removed unnecessary compare
31-Dec-03 RMS Fixed bug in cpu_set_hist
24-Oct-03 RMS Added DMA/DMC support, instruction history
30-Dec-01 RMS Added old PC queue
03-Nov-01 RMS Fixed NOHSA modifier
30-Nov-01 RMS Added extended SET/SHOW support
The register state for the Honeywell 316/516 CPU is:
AR<1:16> A register
BR<1:16> B register
XR<1:16> X register
PC<1:16> P register (program counter)
Y<1:16> memory address register
MB<1:16> memory data register
C overflow flag
EXT extend mode flag
DP double precision mode flag
SC<1:6> shift count
SR[1:4]<0> sense switches 1-4
The Honeywell 316/516 has six instruction formats: memory reference,
I/O, control, shift, skip, and operate.
The memory reference format is:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
|in|xr| op |sc| offset | memory reference
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
<13:10> mnemonic action
0000 (other) see control, shift, skip, operate instructions
0001 JMP P = MA
0010 LDA A = M[MA]
0011 ANA A = A & M[MA]
0100 STA M[MA] = A
0101 ERA A = A ^ M[MA]
0110 ADD A = A + M[MA]
0111 SUB A = A - M[MA]
1000 JST M[MA] = P, P = MA + 1
1001 CAS skip if A == M[MA], double skip if A < M[MA]
1010 IRS M[MA] = M[MA] + 1, skip if M[MA] == 0
1011 IMA A <=> M[MA]
1100 (I/O) see I/O instructions
1101 LDX/STX X = M[MA] (xr = 1), M[MA] = x (xr = 0)
1110 MPY multiply
1111 DIV divide
In non-extend mode, memory reference instructions can access an address
space of 16K words. Multiple levels of indirection are supported, and
each indirect word supplies its own indirect and index bits.
<1,2,7> mode action
0,0,0 sector zero direct MA = IR<8:0>
0,0,1 current direct MA = P<13:9>'IR<8:0>
0,1,0 sector zero indexed MA = IR<8:0> + X
0,1,1 current direct MA = P<13:9>'IR<8:0> + X
1,0,0 sector zero indirect MA = M[IR<8:0>]
1,0,1 current indirect MA = M[P<13:9>'IR<8:0>]
1,1,0 sector zero indirect indexed MA = M[IR<8:0> + X]
1,1,1 current indirect indexed MA = M[MA = P<13:9>'IR<8:0> + X]
In extend mode, memory reference instructions can access an address
space of 32K words. Multiple levels of indirection are supported, but
only post-indexing, based on the original instruction word index flag,
is allowed.
The control format is:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| 0 0 0 0 0 0| opcode | control
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
The shift format is:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| 0 1 0 0 0 0|dr|sz|type | shift count | shift
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| | \-+-/
| | |
| | +--------------------- type
| +------------------------- long/A only
+---------------------------- right/left
The skip format is:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| 1 0 0 0 0 0|rv|po|pe|ev|ze|s1|s2|s3|s4|cz| skip
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| | | | | | | | | |
| | | | | | | | | +- skip if C = 0
| | | | | | | | +---- skip if ssw 4 = 0
| | | | | | | +------- skip if ssw 3 = 0
| | | | | | +---------- skip if ssw 2 = 0
| | | | | +------------- skip if ssw 1 = 0
| | | | +---------------- skip if A == 0
| | | +------------------- skip if A<0> == 0
| | +---------------------- skip if mem par err
| +------------------------- skip if A<15> = 0
+---------------------------- reverse skip sense
The operate format is:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| 1 1 0 0 0 0| opcode | operate
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
The I/O format is:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| op | 1 1 0 0| function | device | I/O transfer
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
The IO transfer instruction controls the specified device.
Depending on the opcode, the instruction may set or clear
the device flag, start or stop I/O, or read or write data.
This routine is the instruction decode routine for the Honeywell
316/516. It is called from the simulator control program to execute
instructions in simulated memory, starting at the simulated PC.
It runs until 'reason' is set non-zero.
General notes:
1. Reasons to stop. The simulator can be stopped by:
HALT instruction
breakpoint encountered
infinite indirection loop
unimplemented instruction and stop_inst flag set
unknown I/O device and stop_dev flag set
I/O error in I/O simulator
2. Interrupts. Interrupts are maintained by parallel variables:
dev_int[2] device interrupt flags
dev_enb[2] device interrupt enable flags
Note that these are actually arrays of two 16 bit words each. The first
word of each vector contains the bits for the standard interrupt devices,
and the second word is the bits for the extended interrupts 1..17. The
IMP uses these extended interrupts, however this was a standard H316 option
and is in no way IMP specific. Actually the H316 supported up to 48 extra
interrupts, but it seems like overkill to implement them all.
In addition, dev_int[0] contains the interrupt enable and interrupt no
defer flags. If interrupt enable and interrupt no defer are set, and
at least one interrupt request is pending, then an interrupt occurs.
The order of flags in these variables corresponds to the order
in the SMK instruction.
3. Non-existent memory. On the H316/516, reads to non-existent memory
return zero, and writes are ignored. In the simulator, the
largest possible memory is instantiated and initialized to zero.
Thus, only writes need be checked against actual memory size.
4. Adding I/O devices. These modules must be modified:
h316_defs.h add interrupt request definition
h316_cpu.c add device dispatch table entry
h316_sys.c add sim_devices table entry
Notes on the behavior of XR:
- XR is "shadowed" by memory location 0 as seen by the program currently
executing. Thus, in extend mode, this is always absolute location 0.
However, if extend mode is off, this is location 0, if the program is
executing in the lower bank, or location 040000, if the program is
executing in the upper bank. Writing XR writes the shadowed memory
location, and vice versa.
- However, the front panel console always equates XR to absolute location
0, regardless of extend mode. There is no direct examine or deposit
to XR; the user must examine or deposit location 0.
*/
#include "h316_defs.h"
#ifdef VM_IMPTIP
#include "h316_imp.h"
#endif
#define PCQ_SIZE 64 /* must be 2**n */
#define PCQ_MASK (PCQ_SIZE - 1)
#define PCQ_ENTRY pcq[pcq_p = (pcq_p - 1) & PCQ_MASK] = PC
#define PCQ_TOP pcq[pcq_p]
#define m7 0001000 /* for generics */
#define m8 0000400
#define m9 0000200
#define m10 0000100
#define m11 0000040
#define m12 0000020
#define m13 0000010
#define m14 0000004
#define m15 0000002
#define m16 0000001
#define HIST_PC 0x40000000
#define HIST_C 0x20000000
#define HIST_EA 0x10000000
#define HIST_MIN 64
#define HIST_MAX 65536
typedef struct {
int32 pc;
int32 ir;
int32 ar;
int32 br;
int32 xr;
int32 ea;
int32 opnd;
t_bool iack; // [RLA] TRUE if an interrupt occurred
} InstHistory;
uint16 M[MAXMEMSIZE] = { 0 }; /* memory */
int32 saved_AR = 0; /* A register */
int32 saved_BR = 0; /* B register */
int32 saved_XR = 0; /* X register */
int32 XR = 0; /* live copy - must be global */
int32 PC = 0; /* P register */
int32 C = 0; /* C register */
int32 ext = 0; /* extend mode */
int32 pme = 0; /* prev mode extend */
int32 extoff_pending = 0; /* extend off pending */
int32 dp = 0; /* double mode */
int32 sc = 0; /* shift count */
int32 ss[4]; /* sense switches */
int32 dev_int = 0; /* dev ready */
int32 dev_enb = 0; /* dev enable */
uint32 ext_ints = 0; // [RLA] 16 if extended interrupts enabled
uint16 dev_ext_int = 0; // [RLA] extended interrupt request bitmap
uint16 dev_ext_enb = 0; // [RLA] extended interrupt enable bitmap
int32 ind_max = 8; /* iadr nest limit */
int32 stop_inst = 1; /* stop on ill inst */
int32 stop_dev = 2; /* stop on ill dev */
uint16 pcq[PCQ_SIZE] = { 0 }; /* PC queue */
int32 pcq_p = 0; /* PC queue ptr */
REG *pcq_r = NULL; /* PC queue reg ptr */
uint32 dma_nch = DMA_MAX; /* number of chan */
uint32 dma_ad[DMA_MAX] = { 0 }; /* DMA addresses */
uint32 dma_wc[DMA_MAX] = { 0 }; /* DMA word count */
uint32 dma_eor[DMA_MAX] = { 0 }; /* DMA end of range */
uint32 chan_req = 0; /* channel requests */
uint32 chan_map[DMA_MAX + DMC_MAX] = { 0 }; /* chan->dev map */
int32 (*iotab[DEV_MAX])(int32 inst, int32 fnc, int32 dat, int32 dev) = { NULL };
int32 hst_p = 0; /* history pointer */
int32 hst_lnt = 0; /* history length */
InstHistory *hst = NULL; /* instruction history */
t_bool devtab_init (void);
int32 dmaio (int32 inst, int32 fnc, int32 dat, int32 dev);
int32 undio (int32 inst, int32 fnc, int32 dat, int32 dev);
t_stat cpu_ex (t_value *vptr, t_addr addr, UNIT *uptr, int32 sw);
t_stat cpu_dep (t_value val, t_addr addr, UNIT *uptr, int32 sw);
t_stat cpu_reset (DEVICE *dptr);
t_stat cpu_set_noext (UNIT *uptr, int32 val, CONST char *cptr, void *desc);
t_stat cpu_set_size (UNIT *uptr, int32 val, CONST char *cptr, void *desc);
t_stat cpu_set_hist (UNIT *uptr, int32 val, CONST char *cptr, void *desc);
t_stat cpu_show_hist (FILE *st, UNIT *uptr, int32 val, CONST void *desc);
t_stat cpu_show_dma (FILE *st, UNIT *uptr, int32 val, CONST void *desc);
t_stat cpu_set_nchan (UNIT *uptr, int32 val, CONST char *cptr, void *desc);
t_stat cpu_show_nchan (FILE *st, UNIT *uptr, int32 val, CONST void *desc);
t_stat cpu_set_interrupts (UNIT *uptr, int32 val, CONST char *cptr, void *desc);
t_stat cpu_show_interrupts (FILE *st, UNIT *uptr, int32 val, CONST void *desc);
int32 sim_ota_2024 (int32 inst, int32 fnc, int32 dat, int32 dev);
int32 cpu_interrupt (int32 vec);
int32 cpu_ext_interrupt (void);
/* CPU data structures
cpu_dev CPU device descriptor
cpu_unit CPU unit descriptor
cpu_reg CPU register list
cpu_mod CPU modifiers list
*/
DIB cpu_dib = { DMA, 1, IOBUS, IOBUS, INT_V_NONE, INT_V_NONE, &dmaio, 0 };
UNIT cpu_unit = {
UDATA (NULL, UNIT_FIX+UNIT_BINK+UNIT_EXT+UNIT_HSA+UNIT_DMC, MAXMEMSIZE)
};
REG cpu_reg[] = {
{ ORDATA (P, PC, 15) },
{ ORDATA (A, saved_AR, 16) },
{ ORDATA (B, saved_BR, 16) },
{ ORDATA (X, saved_XR, 16) },
{ ORDATA (SC, sc, 16) },
{ FLDATA (C, C, 0) },
{ FLDATA (EXT, ext, 0) },
{ FLDATA (PME, pme, 0) },
{ FLDATA (EXT_OFF, extoff_pending, 0) },
{ FLDATA (DP, dp, 0) },
{ FLDATA (SS1, ss[0], 0) },
{ FLDATA (SS2, ss[1], 0) },
{ FLDATA (SS3, ss[2], 0) },
{ FLDATA (SS4, ss[3], 0) },
{ FLDATA (ION, dev_int, INT_V_ON) },
{ FLDATA (INODEF, dev_int, INT_V_NODEF) },
{ FLDATA (START, dev_int, INT_V_START) },
{ ORDATA (DEVINT, dev_int, 16), REG_RO },
{ ORDATA (DEVENB, dev_enb, 16), REG_RO },
{ ORDATA (EXTINT, dev_ext_int, 16), REG_RO },
{ ORDATA (EXTENB, dev_ext_enb, 16), REG_RO },
{ ORDATA (CHREQ, chan_req, DMA_MAX + DMC_MAX) },
{ BRDATA (DMAAD, dma_ad, 8, 16, DMA_MAX) },
{ BRDATA (DMAWC, dma_wc, 8, 16, DMA_MAX) },
{ BRDATA (DMAEOR, dma_eor, 8, 1, DMA_MAX) },
{ ORDATA (DMANCH, dma_nch, 3), REG_HRO },
{ FLDATA (MPERDY, dev_int, INT_V_MPE) },
{ FLDATA (MPEENB, dev_enb, INT_V_MPE) },
{ FLDATA (STOP_INST, stop_inst, 0) },
{ FLDATA (STOP_DEV, stop_dev, 1) },
{ DRDATA (INDMAX, ind_max, 8), REG_NZ + PV_LEFT },
{ BRDATA (PCQ, pcq, 8, 15, PCQ_SIZE), REG_RO + REG_CIRC },
{ ORDATA (PCQP, pcq_p, 6), REG_HRO },
{ ORDATA (WRU, sim_int_char, 8) },
{ NULL }
};
MTAB cpu_mod[] = {
{ UNIT_EXT, 0, "no extend", "NOEXTEND", &cpu_set_noext },
{ UNIT_EXT, UNIT_EXT, "extend", "EXTEND", NULL },
{ UNIT_HSA, 0, "no HSA", "NOHSA", NULL },
{ UNIT_HSA, UNIT_HSA, "HSA", "HSA", NULL },
{ UNIT_MSIZE, 4096, NULL, "4K", &cpu_set_size },
{ UNIT_MSIZE, 8192, NULL, "8K", &cpu_set_size },
{ UNIT_MSIZE, 12288, NULL, "12K", &cpu_set_size },
{ UNIT_MSIZE, 16384, NULL, "16K", &cpu_set_size },
{ UNIT_MSIZE, 24576, NULL, "24K", &cpu_set_size },
{ UNIT_MSIZE, 32768, NULL, "32K", &cpu_set_size },
{ MTAB_XTD | MTAB_VDV, 0, "channels", "CHANNELS",
&cpu_set_nchan, &cpu_show_nchan, NULL },
{ MTAB_XTD | MTAB_VDV, 0, NULL, "DMA", // [RLA] this is the way it's
&cpu_set_nchan, NULL, NULL }, // [RLA] documented to work!
{ UNIT_DMC, 0, "no DMC", "NODMC", NULL },
{ UNIT_DMC, UNIT_DMC, "DMC", "DMC", NULL },
{ MTAB_XTD|MTAB_VDV|MTAB_NMO|MTAB_SHP, 0, "HISTORY", "HISTORY",
&cpu_set_hist, &cpu_show_hist },
{ MTAB_XTD | MTAB_VDV, 0, "extended interrupts", "EXTINT",
&cpu_set_interrupts, &cpu_show_interrupts, NULL },
{ 0 }
};
DEVICE cpu_dev = {
"CPU", &cpu_unit, cpu_reg, cpu_mod,
1, 8, 15, 1, 8, 16,
&cpu_ex, &cpu_dep, &cpu_reset,
NULL, NULL, NULL,
&cpu_dib, 0
};
t_stat sim_instr (void)
{
int32 AR, BR, MB, Y, t1, t2, t3, skip, dev;
uint32 ut;
t_bool iack; // [RLA] TRUE if an interrupt was taken this cycle
t_stat reason;
t_stat Ea (int32 inst, int32 *addr);
t_stat Write (int32 addr, int32 val); // [RLA] Write() can now cause a break
int32 Add16 (int32 val1, int32 val2);
int32 Add31 (int32 val1, int32 val2);
int32 Operate (int32 MB, int32 AR);
#define Read(ad) M[(ad)]
#define GETDBL_S(h,l) (((h) << 15) | ((l) & MMASK))
#define GETDBL_U(h,l) (((h) << 16) | (l))
#define PUTDBL_S(x) AR = ((x) >> 15) & DMASK; \
BR = (BR & SIGN) | ((x) & MMASK)
#define PUTDBL_U(x) AR = ((x) >> 16) & DMASK; \
BR = (x) & DMASK
#define PUTDBL_Z(x) AR = ((x) >> 15) & DMASK; \
BR = (x) & MMASK
#define SEXT(x) (((x) & SIGN)? ((x) | ~DMASK): ((x) & DMASK))
#define NEWA(c,n) (ext? (((c) & ~X_AMASK) | ((n) & X_AMASK)): \
(((c) & ~NX_AMASK) | ((n) & NX_AMASK)))
/* Restore register state */
if (devtab_init ()) /* init tables */
return SCPE_STOP;
AR = saved_AR & DMASK; /* restore reg */
BR = saved_BR & DMASK;
XR = saved_XR & DMASK;
PC = PC & ((cpu_unit.flags & UNIT_EXT)? X_AMASK: NX_AMASK); /* mask PC */
reason = 0;
/* Main instruction fetch/decode loop */
while (reason == 0) { /* loop until halted */
if (sim_interval <= 0) { /* check clock queue */
if ((reason = sim_process_event ()))
break;
}
/* Channel breaks (DMA and DMC) */
if (chan_req) { /* channel request? */
int32 i, t, ch, dev, st, end, ad, dmcad;
t_stat r;
for (i = 0, ch = chan_req; ch != 0; i++, ch = ch >> 1) {
if (ch & 1) { /* req on chan i? */
dev = chan_map[i]; /* get dev for chan */
if (iotab[dev] == &undio)
return SCPE_IERR;
chan_req = chan_req & ~(1 << i); /* clear req */
if (Q_DMA (i)) /* DMA? */
st = dma_ad[i];
else { /* DMC */
dmcad = DMC_BASE + ((i - DMC_V_DMC1) << 1);
st = Read (dmcad); /* DMC ctrl word */
}
ad = st & X_AMASK; /* get curr addr */
if (st & DMA_IN) { /* input? */
t = iotab[dev] (ioINA, 0, 0, dev); /* input word */
if ((t & IOT_SKIP) == 0)
return STOP_DMAER;
if ((r = t >> IOT_V_REASON) != 0)
return r;
// [RLA] Note that we intentionally ignore address breaks here!
Write (ad, t & DMASK); /* write to mem */
}
else { /* no, output */
t = iotab[dev] (ioOTA, 0, Read (ad), dev); /* output word */
if ((t & IOT_SKIP) == 0)
return STOP_DMAER;
if ((r = (t >> IOT_V_REASON)))
return r;
}
if (Q_DMA (i)) { /* DMA? */
dma_ad[i] = (dma_ad[i] & DMA_IN) | ((ad + 1) & X_AMASK);
dma_wc[i] = (dma_wc[i] + 1) & 077777; /* update wc */
if (dma_wc[i] == 0) { /* done? */
dma_eor[i] = 1; /* set end of range */
t = iotab[dev] (ioEND, 0, 0, dev); /* send end range */
if ((r = t >> IOT_V_REASON) != 0)
return r;
}
}
else { /* DMC */
st = (st & DMA_IN) | ((ad + 1) & X_AMASK);
// [RLA] Note that we intentionally ignore address breaks here!
Write (dmcad, st); /* update start */
end = Read (dmcad + 1); /* get end */
if (((ad ^ end) & X_AMASK) == 0) { /* start == end? */
t = iotab[dev] (ioEND, 0, 0, dev); /* send end range */
if ((r = t >> IOT_V_REASON) != 0)
return r;
} /* end if end range */
} /* end else DMC */
} /* end if chan i */
} /* end for */
} /* end if chan_req */
/* Interrupts */
//[RLA] Todo - add WDT interrupts ????
iack = FALSE;
if ((dev_int & (INT_PEND|INT_NMI|dev_enb)) > INT_PEND) { // [RLA] check for standard interrupt
MB = cpu_interrupt(M_INT);
iack = TRUE;
}
else if (((dev_ext_int & dev_ext_enb) != 0) // [RLA] check for extended interrupt
&& ((dev_int & INT_PEND) == INT_PEND)) {
MB = cpu_ext_interrupt ();
iack = TRUE;
}
/* Instruction fetch */
else {
if (sim_brk_summ &&
sim_brk_test (PC, SWMASK ('E'))) { /* breakpoint? */
reason = STOP_IBKPT; /* stop simulation */
break;
}
Y = PC; /* set mem addr */
MB = Read (Y); /* fetch instr */
PC = NEWA (Y, Y + 1); /* incr PC */
dev_int = dev_int | INT_NODEF;
}
dev_int = dev_int & ~INT_START; /* clr start button int */
sim_interval = sim_interval - 1;
if (hst_lnt) { /* instr hist? */
hst_p = (hst_p + 1); /* next entry */
if (hst_p >= hst_lnt)
hst_p = 0;
hst[hst_p].pc = Y | HIST_PC | (C? HIST_C: 0); /* fill slots */
hst[hst_p].ir = MB;
hst[hst_p].ar = AR;
hst[hst_p].br = BR;
hst[hst_p].xr = XR;
hst[hst_p].iack = iack; // [RLA] record if interrupt taken
}
/* Memory reference instructions */
switch (I_GETOP (MB)) { /* case on <1:6> */
case 001: case 021: case 041: case 061: /* JMP */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
PCQ_ENTRY; /* save PC */
PC = NEWA (PC, Y); /* set new PC */
if (extoff_pending) /* cond ext off */
ext = extoff_pending = 0;
break;
case 002: case 022: case 042: case 062: /* LDA */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
if (dp) { /* double prec? */
AR = Read (Y & ~1); /* get doubleword */
BR = Read (Y | 1);
sc = 0;
}
else AR = Read (Y); /* no, get word */
break;
case 003: case 023: case 043: case 063: /* ANA */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
AR = AR & Read (Y);
break;
case 004: case 024: case 044: case 064: /* STA */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
if ((reason = Write(Y, AR)))
break; /* [RLA] store A */
if (dp) { /* double prec? */
if ((reason = Write(Y | 1, BR)))
break; /* [RLA] store B */
sc = 0;
}
break;
case 005: case 025: case 045: case 065: /* ERA */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
AR = AR ^ Read (Y);
break;
case 006: case 026: case 046: case 066: /* ADD */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
if (dp) { /* double prec? */
t1 = GETDBL_S (AR, BR); /* get A'B */
t2 = GETDBL_S (Read (Y & ~1), Read (Y | 1));
t1 = Add31 (t1, t2); /* 31b add */
PUTDBL_Z (t1);
sc = 0;
}
else AR = Add16 (AR, Read (Y)); /* no, 16b add */
break;
case 007: case 027: case 047: case 067: /* SUB */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
if (dp) { /* double prec? */
t1 = GETDBL_S (AR, BR); /* get A'B */
t2 = GETDBL_S (Read (Y & ~1), Read (Y | 1));
t1 = Add31 (t1, -t2); /* 31b sub */
PUTDBL_Z (t1);
sc = 0;
}
else AR = Add16 (AR, (-Read (Y)) & DMASK); /* no, 16b sub */
break;
case 010: case 030: case 050: case 070: /* JST */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
MB = NEWA (Read (Y), PC); /* merge old PC */
if ((reason = Write(Y, MB)))
break; // [RLA]
PCQ_ENTRY;
PC = NEWA (PC, Y + 1); /* set new PC */
break;
case 011: case 031: case 051: case 071: /* CAS */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
MB = Read (Y);
if (AR == MB)
PC = NEWA (PC, PC + 1);
else if (SEXT (AR) < SEXT (MB))
PC = NEWA (PC, PC + 2);
break;
case 012: case 032: case 052: case 072: /* IRS */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
MB = (Read (Y) + 1) & DMASK; /* incr, rewrite */
if ((reason = Write(Y, MB)))
break; // [RLA]
if (MB == 0) /* skip if zero */
PC = NEWA (PC, PC + 1);
break;
case 013: case 033: case 053: case 073: /* IMA */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
MB = Read (Y);
if ((reason = Write(Y, AR)))
break; /* [RLA] A to mem */
AR = MB; /* mem to A */
break;
case 015: case 055: /* STX */
if ((reason = Ea (MB & ~IDX, &Y))) /* eff addr */
break;
if ((reason = Write(Y, XR)))
break; /* [RLA] store XR */
break;
case 035: case 075: /* LDX */
if ((reason = Ea (MB & ~IDX, &Y))) /* eff addr */
break;
XR = Read (Y); /* load XR */
M[M_XR] = XR; /* update mem too */
break;
case 016: case 036: case 056: case 076: /* MPY */
if (cpu_unit.flags & UNIT_HSA) { /* installed? */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
t1 = SEXT (AR) * SEXT (Read (Y));
PUTDBL_Z (t1);
sc = 0;
}
else reason = stop_inst;
break;
case 017: case 037: case 057: case 077: /* DIV */
if (cpu_unit.flags & UNIT_HSA) { /* installed? */
if ((reason = Ea (MB, &Y))) /* eff addr */
break;
t2 = SEXT (Read (Y)); /* divr */
if (t2) { /* divr != 0? */
t1 = GETDBL_S (SEXT (AR), BR); /* get A'B signed */
BR = (t1 % t2) & DMASK; /* remainder */
t1 = t1 / t2; /* quotient */
AR = t1 & DMASK;
if ((t1 > MMASK) || (t1 < (-SIGN)))
C = 1;
else C = 0;
sc = 0;
}
else C = 1;
}
else reason = stop_inst;
break;
/* I/O instructions */
case 014: /* OCP */
dev = MB & DEVMASK;
t2 = iotab[dev] (ioOCP, I_GETFNC (MB), AR, dev);
reason = t2 >> IOT_V_REASON;
break;
case 034: /* SKS */
dev = MB & DEVMASK;
t2 = iotab[dev] (ioSKS, I_GETFNC (MB), AR, dev);
reason = t2 >> IOT_V_REASON;
if (t2 & IOT_SKIP) /* skip? */
PC = NEWA (PC, PC + 1);
break;
case 054: /* INA */
dev = MB & DEVMASK;
if (MB & INCLRA)
AR = 0;
t2 = iotab[dev] (ioINA, I_GETFNC (MB & ~INCLRA), AR, dev);
reason = t2 >> IOT_V_REASON;
if (t2 & IOT_SKIP) /* skip? */
PC = NEWA (PC, PC + 1);
AR = t2 & DMASK; /* data */
break;
case 074: /* OTA */
dev = MB & DEVMASK;
// [RLA] OTA w/devices 20 or 24 are SMK or OTK!
if ((dev == 020) || (dev == 024))
t2 = sim_ota_2024(ioOTA, I_GETFNC (MB), AR, dev);
else
t2 = iotab[dev] (ioOTA, I_GETFNC (MB), AR, dev);
reason = t2 >> IOT_V_REASON;
if (t2 & IOT_SKIP) /* skip? */
PC = NEWA (PC, PC + 1);
break;
/* Control */
case 000:
if ((MB & 1) == 0) { /* HLT */
if ((reason = sim_process_event ()) != SCPE_OK)
break;
reason = STOP_HALT;
break;
}
if (MB & m14) { /* SGL, DBL */
if (cpu_unit.flags & UNIT_HSA)
dp = (MB & m15)? 1: 0;
else reason = stop_inst;
}
if (MB & m13) { /* DXA, EXA */
if (!(cpu_unit.flags & UNIT_EXT))
reason = stop_inst;
else if (MB & m15) { /* EXA */
ext = 1;
extoff_pending = 0; /* DXA */
}
else extoff_pending = 1;
}
if (MB & m12) /* RMP */
CLR_INT (INT_MPE);
if (MB & m11) { /* SCA, INK */
if (MB & m15) /* INK */
AR = (C << 15) | (dp << 14) | (pme << 13) | (sc & 077);
else if (cpu_unit.flags & UNIT_HSA) /* SCA */
AR = sc & 077;
else reason = stop_inst;
}
else if (MB & m10) { /* NRM */
if (cpu_unit.flags & UNIT_HSA) {
for (sc = 0;
(sc < 32) && ((AR & SIGN) == ((AR << 1) & SIGN));
sc++) {
AR = (AR & SIGN) | ((AR << 1) & MMASK) |
((BR >> 14) & 1);
BR = (BR & SIGN) | ((BR << 1) & MMASK);
}
sc = sc & 037;
}
else reason = stop_inst;
}
else if (MB & m9) { /* IAB */
sc = BR;
BR = AR;
AR = sc;
}
if (MB & m8) /* ENB */
dev_int = (dev_int | INT_ON) & ~INT_NODEF;
if (MB & m7) /* INH */
dev_int = dev_int & ~INT_ON;
break;
/* Shift
Shifts are microcoded as follows:
op<7> = right/left
op<8> = long/short
op<9> = shift/rotate (rotate bits "or" into new position)
op<10> = logical/arithmetic
If !op<7> && op<10> (right arithmetic), A<1> propagates rightward
If op<7> && op<10> (left arithmetic), C is set if A<1> changes state
If !op<8> && op<10> (long arithmetic), B<1> is skipped
This microcoding "explains" how the 4 undefined opcodes actually work
003 = long arith rotate right, skip B<1>, propagate A<1>,
bits rotated out "or" into A<1>
007 = short arith rotate right, propagate A<1>,
bits rotated out "or" into A<1>
013 = long arith rotate left, skip B<1>, C = overflow
017 = short arith rotate left, C = overflow
*/
case 020:
C = 0; /* clear C */
sc = 0; /* clear sc */
if ((t1 = (-MB) & SHFMASK) == 0) /* shift count */
break;
switch (I_GETFNC (MB)) { /* case shift fnc */
case 000: /* LRL */
if (t1 > 32) /* >32? all 0 */
ut = 0;
else {
ut = GETDBL_U (AR, BR); /* get A'B */
C = (ut >> (t1 - 1)) & 1; /* C = last out */
if (t1 == 32) /* =32? all 0 */
ut = 0;
else ut = ut >> t1; /* log right */
}
PUTDBL_U (ut); /* store A,B */
break;
case 001: /* LRS */
if (t1 > 31) /* limit to 31 */
t1 = 31;
t2 = GETDBL_S (SEXT (AR), BR); /* get A'B signed */
C = (t2 >> (t1 - 1)) & 1; /* C = last out */
t2 = t2 >> t1; /* arith right */
PUTDBL_S (t2); /* store A,B */
break;
case 002: /* LRR */
t2 = t1 % 32; /* mod 32 */
ut = GETDBL_U (AR, BR); /* get A'B */
ut = (ut >> t2) | (ut << (32 - t2)); /* rot right */
C = (ut >> 31) & 1; /* C = A<1> */
PUTDBL_U (ut); /* store A,B */
break;
case 003: /* "long right arot" */
if ((reason = stop_inst)) /* stop on undef? */
break;
for (t2 = 0; t2 < t1; t2++) { /* bit by bit */
C = BR & 1; /* C = last out */
BR = (BR & SIGN) | ((AR & 1) << 14) |
((BR & MMASK) >> 1);
AR = ((AR & SIGN) | (C << 15)) | (AR >> 1);
}
break;
case 004: /* LGR */
if (t1 > 16) /* > 16? all 0 */
AR = 0;
else {
C = (AR >> (t1 - 1)) & 1; /* C = last out */
AR = (AR >> t1) & DMASK; /* log right */
}
break;
case 005: /* ARS */
if (t1 > 16) /* limit to 16 */
t1 = 16;
C = ((SEXT (AR)) >> (t1 - 1)) & 1; /* C = last out */
AR = ((SEXT (AR)) >> t1) & DMASK; /* arith right */
break;
case 006: /* ARR */
t2 = t1 % 16; /* mod 16 */
AR = ((AR >> t2) | (AR << (16 - t2))) & DMASK;
C = (AR >> 15) & 1; /* C = A<1> */
break;
case 007: /* "short right arot" */
if ((reason = stop_inst)) /* stop on undef? */
break;
for (t2 = 0; t2 < t1; t2++) { /* bit by bit */
C = AR & 1; /* C = last out */
AR = ((AR & SIGN) | (C << 15)) | (AR >> 1);
}
break;
case 010: /* LLL */
if (t1 > 32) /* > 32? all 0 */
ut = 0;
else {
ut = GETDBL_U (AR, BR); /* get A'B */
C = (ut >> (32 - t1)) & 1; /* C = last out */
if (t1 == 32) /* =32? all 0 */
ut = 0;
else ut = ut << t1; /* log left */
}
PUTDBL_U (ut); /* store A,B */
break;
case 011: /* LLS */
if (t1 > 31) /* limit to 31 */
t1 = 31;
t2 = GETDBL_S (SEXT (AR), BR); /* get A'B */
t3 = t2 << t1; /* "arith" left */
PUTDBL_S (t3); /* store A'B */
if ((t2 >> (31 - t1)) != ((AR & SIGN)? -1: 0)) /* shf out = sgn? */
C = 1;
break;
case 012: /* LLR */
t2 = t1 % 32; /* mod 32 */
ut = GETDBL_U (AR, BR); /* get A'B */
ut = (ut << t2) | (ut >> (32 - t2)); /* rot left */
C = ut & 1; /* C = B<16> */
PUTDBL_U (ut); /* store A,B */
break;
case 013: /* "long left arot" */
if ((reason = stop_inst)) /* stop on undef? */
break;
for (t2 = 0; t2 < t1; t2++) { /* bit by bit */
AR = (AR << 1) | ((BR >> 14) & 1);
BR = (BR & SIGN) | ((BR << 1) & MMASK) |
((AR >> 16) & 1);
if ((AR & SIGN) != ((AR >> 1) & SIGN)) C = 1;
AR = AR & DMASK;
}
break;
case 014: /* LGL */
if (t1 > 16) /* > 16? all 0 */
AR = 0;
else {
C = (AR >> (16 - t1)) & 1; /* C = last out */
AR = (AR << t1) & DMASK; /* log left */
}
break;
case 015: /* ALS */
if (t1 > 16) /* limit to 16 */
t1 = 16;
t2 = SEXT (AR); /* save AR */
AR = (AR << t1) & DMASK; /* "arith" left */
if ((t2 >> (16 - t1)) != /* shf out + sgn */
((AR & SIGN)? -1: 0)) C = 1;
break;
case 016: /* ALR */
t2 = t1 % 16; /* mod 16 */
AR = ((AR << t2) | (AR >> (16 - t2))) & DMASK;
C = AR & 1; /* C = A<16> */
break;
case 017: /* "short left arot" */
if ((reason = stop_inst)) /* stop on undef? */
break;
for (t2 = 0; t2 < t1; t2++) { /* bit by bit */
if ((AR & SIGN) != ((AR << 1) & SIGN)) C = 1;
AR = ((AR << 1) | (AR >> 15)) & DMASK;
}
break; /* end case fnc */
}
break;
/* Skip */
case 040:
skip = 0;
if (((MB & 000001) && C) || /* SSC */
((MB & 000002) && ss[3]) || /* SS4 */
((MB & 000004) && ss[2]) || /* SS3 */
((MB & 000010) && ss[1]) || /* SS2 */
((MB & 000020) && ss[0]) || /* SS1 */
((MB & 000040) && AR) || /* SNZ */
((MB & 000100) && (AR & 1)) || /* SLN */
((MB & 000200) && (TST_INTREQ (INT_MPE))) || /* SPS */
((MB & 000400) && (AR & SIGN))) /* SMI */
skip = 1;
if ((MB & 001000) == 0) /* reverse? */
skip = skip ^ 1;
PC = NEWA (PC, PC + skip);
break;
/* Operate */
case 060:
if (MB == 0140024) /* CHS */
AR = AR ^ SIGN;
else if (MB == 0140040) /* CRA */
AR = 0;
else if (MB == 0140100) /* SSP */
AR = AR & ~SIGN;
else if (MB == 0140200) /* RCB */
C = 0;
else if (MB == 0140320) { /* CSA */
C = (AR & SIGN) >> 15;
AR = AR & ~SIGN;
}
else if (MB == 0140401) /* CMA */
AR = AR ^ DMASK;
else if (MB == 0140407) { /* TCA */
AR = (-AR) & DMASK;
sc = 0;
}
else if (MB == 0140500) /* SSM */
AR = AR | SIGN;
else if (MB == 0140600) /* SCB */
C = 1;
else if (MB == 0141044) /* CAR */
AR = AR & 0177400;
else if (MB == 0141050) /* CAL */
AR = AR & 0377;
else if (MB == 0141140) /* ICL */
AR = AR >> 8;
else if (MB == 0141206) /* AOA */
AR = Add16 (AR, 1);
else if (MB == 0141216) /* ACA */
AR = Add16 (AR, C);
else if (MB == 0141240) /* ICR */
AR = (AR << 8) & DMASK;
else if (MB == 0141340) /* ICA */
AR = ((AR << 8) | (AR >> 8)) & DMASK;
else if ((reason = stop_inst))
break;
else AR = Operate (MB, AR); /* undefined */
break;
} /* end case op */
} /* end while */
saved_AR = AR & DMASK;
saved_BR = BR & DMASK;
saved_XR = XR & DMASK;
pcq_r->qptr = pcq_p; /* update pc q ptr */
return reason;
}
/* Effective address
The effective address calculation consists of three phases:
- base address calculation: 0/pagenumber'displacement
- (extend): indirect address resolution
(non-extend): pre-indexing
- (extend): post-indexing
(non-extend): indirect address/post-indexing resolution
In extend mode, address calculations are carried out to 16b
and masked to 15b at exit. In non-extend mode, address bits
<1:2> are preserved by the NEWA macro; address bit <1> is
masked at exit.
*/
t_stat Ea (int32 IR, int32 *addr)
{
int32 i;
int32 Y = IR & (IA | DISP); /* ind + disp */
if (IR & SC) Y = ((PC - 1) & PAGENO) | Y; /* cur sec? + pageno */
if (ext) { /* extend mode? */
for (i = 0; (i < ind_max) && (Y & IA); i++) { /* resolve ind addr */
Y = Read (Y & X_AMASK); /* get ind addr */
}
if (IR & IDX) Y = Y + XR; /* post-index */
} /* end if ext */
else { /* non-extend */
Y = NEWA (PC, Y + ((IR & IDX)? XR: 0)); /* pre-index */
for (i = 0; (i < ind_max) && (IR & IA); i++) { /* resolve ind addr */
IR = Read (Y & X_AMASK); /* get ind addr */
Y = NEWA (Y, IR + ((IR & IDX)? XR: 0)); /* post-index */
}
} /* end else */
*addr = Y = Y & X_AMASK; /* return addr */
if (hst_lnt) { /* history? */
hst[hst_p].pc = hst[hst_p].pc | HIST_EA;
hst[hst_p].ea = Y;
hst[hst_p].opnd = Read (Y);
}
if (i >= ind_max)
return STOP_IND; /* too many ind? */
return SCPE_OK;
}
/* Write memory */
t_stat Write (int32 addr, int32 val)
{
// [RLA] Write() now checks for address breaks ...
if (((addr == 0) || (addr >= 020)) && MEM_ADDR_OK (addr))
M[addr] = val;
if (addr == M_XR) /* write XR loc? */
XR = val;
// [RLA] Implement "break on memory write" ...
if (sim_brk_summ && sim_brk_test (addr, SWMASK ('W')))
return STOP_IBKPT;
else
return SCPE_OK;
}
/* Add */
int32 Add16 (int32 v1, int32 v2)
{
int32 r = v1 + v2;
if (((v1 ^ ~v2) & (v1 ^ r)) & SIGN)
C = 1;
else C = 0;
return (r & DMASK);
}
int32 Add31 (int32 v1, int32 v2)
{
int32 r = v1 + v2;
if (((v1 ^ ~v2) & (v1 ^ r)) & DP_SIGN)
C = 1;
else C = 0;
return r;
}
// [RLA] Standard (fixed vector) interrupt action ...
int32 cpu_interrupt (int32 vec)
{
pme = ext; /* save extend */
if (cpu_unit.flags & UNIT_EXT)
ext = 1; /* ext opt? extend on */
dev_int = dev_int & ~INT_ON; /* intr off */
return 0120000 | vec; /* inst = JST* vector */
}
// [RLA] Extended (priority) interrupt action ...
int32 cpu_ext_interrupt (void) {
// Unlike the standard interrupts, which have a fixed vector shared by all
// devices, the extended interrupts have a unique vector for every device.
// Moreover, extended interrupts are prioritized so that the lowest numbered
// interrupts have priority. That means we have to actually scan the bitmap
// of active interrupts to figure out which one to take.
//
// One uncomfortable thing about the external interrupts is that it appears
// that they were edge triggered - once an interrupt on a given level was
// granted, that interrupt wouldn't occur again until another edge occurred on
// the same request. I'm "uncomfortable" with this because it's different from
// the way the standard interrupt works - that's completely level sensitive.
// Still, this Honeywell document
//
// http://bitsavers.informatik.uni-stuttgart.de/pdf/honeywell/series16/h316/70130072167D_316_Interfacing_Apr73.pdf
//
// (read Chapter 4, Priority Interrupts, the very first paragraph) at least
// seems to imply edge triggering. And the IMP firmware is written as if they
// are edge triggered - there are many cases (modem output, task, RTC) where
// the IMP code does nothing to clear the interrupt request flag. So we're
// going with edge triggered version for now...
int32 i; uint16 m, irq;
irq = dev_ext_int & dev_ext_enb;
for (i = 1, m = SIGN; m != 0; ++i, m >>= 1) {
if ((irq & m) != 0) {
// Extended interrupts are edge triggered (see above) - when this
// interrupt is granted, clear the request ...
CLR_EXT_INT(m);
return cpu_interrupt(M_INT+i);
}
}
// If we get here, it means that we were called with no interrupt bits set.
// That really should never happen, so just HALT ...
return(0);
}
/* Unimplemented I/O device */
int32 undio (int32 op, int32 fnc, int32 val, int32 dev)
{
return ((stop_dev << IOT_V_REASON) | val);
}
/* [RLA] Special I/O devices */
int32 sim_ota_2024 (int32 inst, int32 fnc, int32 dat, int32 dev)
{
// OTA instructions with a device code of 20 or 24 are really SMK
// (Set interrupt Mask) instructions. OTA 20 sets the standard H316
// interrupt mask, and OTA 120, OTA 220 and OTA 320 set the extended
// interrupt mask (of which only one, OTA 120, is used by the IMP).
//
// Further, OTA 1020 is the OTK instruction which sets special CPU
// flags (single or double precision HSA, extended addressing mode,
// the carry flag, etc).
//
// This routine implements these special OTKs as part of the CPU.
// That allows us to implement the extra interrupt masks needed by the
// IMP, and it also allows the CLK device to be disabled without losing
// the SMK or OTK instructions. The clock was an option on the original
// H316 and is not required to be present, and the IMP in particular
// needs it to be disabled.
// Although OTA 24 is reserved nothing we currently simulate uses it!
if (dev == 024)
return IOBADFNC (dat);
// Device code 20...
switch (fnc) {
case 000: // SMK 020 - set standard interrupt mask
dev_enb = dat;
break;
case 001: // SMK 120 - set extended interrupt mask #1
if (ext_ints < 16)
return IOBADFNC(dat);
dev_ext_enb = dat;
break;
case 002: // SMK 220 - set extended interrupt mask #2
case 003: // SMK 320 - set extended interrupt mask #3
return IOBADFNC(dat);
case 010: // OTK - output keys
C = (dat >> 15) & 1; /* set C */
if (cpu_unit.flags & UNIT_HSA) /* HSA included? */
dp = (dat >> 14) & 1; /* set dp */
if (cpu_unit.flags & UNIT_EXT) { /* ext opt? */
if (dat & 020000) { /* ext set? */
ext = 1; /* yes, set */
extoff_pending = 0;
}
else extoff_pending = 1; /* no, clr later */
}
sc = dat & 037; /* set sc */
break;
default:
return IOBADFNC (dat);
}
return dat;
}
/* DMA control */
int32 dmaio (int32 inst, int32 fnc, int32 dat, int32 dev)
{
int32 ch = (fnc - 1) & 03;
switch (inst) { /* case on opcode */
case ioOCP: /* OCP */
if ((fnc >= 001) && (fnc <= 004)) { /* load addr ctr */
dma_ad[ch] = dat;
dma_wc[ch] = 0;
dma_eor[ch] = 0;
}
else if ((fnc >= 011) && (fnc <= 014)) /* load range ctr */
dma_wc[ch] = (dma_wc[ch] | dat) & 077777;
else return IOBADFNC (dat); /* undefined */
break;
case ioINA: /* INA */
if ((fnc >= 011) && (fnc <= 014)) {
if (dma_eor[ch]) /* end range? nop */
return dat;
return IOSKIP (0100000 | dma_wc[ch]); /* return range */
}
else return IOBADFNC (dat);
}
return dat;
}
/* Undefined operate instruction. This code is reached when the
opcode does not correspond to a standard operate instruction.
It simulates the behavior of the actual logic.
An operate instruction executes in 4 or 6 phases. A 'normal'
instruction takes 4 phases:
t1 t1
t2/tlate t2/t2 extended into t3
t3/tlate t3
t4 t4
A '1.5 cycle' instruction takes 6 phases:
t1 t1
t2/tlate t2/t2 extended into t3
t3/tlate t3
t2/tlate 'special' t2/t2 extended into t3
t3/tlate t3
t4 t4
The key signals, by phase, are the following
tlate EASTL enable A to sum leg 1 (else 0)
(((m12+m16)x!azzzz)+(m9+m11+azzzz)
EASBM enable 0 to sum leg 2 (else 177777)
(m9+m11+azzzz)
JAMKN jam carry network to 0 = force XOR
((m12+m16)x!azzzz)
EIKI7 force carry into adder
((m15x(C+!m13))x!JAMKN)
t3 CLDTR set D to 177777 (always)
ESDTS enable adder sum to D (always)
SETAZ enable repeat cycle = set azzzz
(m8xm15)
if azzzz {
t2 CLATR clear A register (due to azzzz)
EDAHS enable D high to A high register (due to azzzz)
EDALS enable D low to A low register (due to azzzz)
tlate, t3 as above
}
t4 CLATR clear A register
(m11+m15+m16)
CLA1R clear A1 register
(m10+m14)
EDAHS enable D high to A high register
((m11xm14)+m15+m16)
EDALS enable D low to A low register
((m11xm13)+m15+m16)
ETAHS enable D transposed to A high register
(m9xm11)
ETALS enable D transposed to A low register
(m10xm11)
EDA1R enable D1 to A1 register
((m8xm10)+m14)
CBITL clear C, conditionally set C from adder output
(m9x!m11)
CBITG conditionally set C if D1
(m10xm12xD1)
CBITE unconditionally set C
(m8xm9)
*/
int32 Operate (int32 MB, int32 AR)
{
int32 D, jamkn, eiki7, easbm, eastl, setaz;
int32 clatr, cla1r, edahs, edals, etahs, etals, eda1r;
int32 cbitl, cbitg, cbite;
int32 aleg, bleg, ARx;
/* Phase tlate */
ARx = AR; /* default */
jamkn = (MB & (m12+m16)) != 0; /* m12+m16 */
easbm = (MB & (m9+m11)) != 0; /* m9+m11 */
eastl = jamkn || easbm; /* m9+m11+m12+m16 */
setaz = (MB & (m8+m15)) == (m8+m15); /* m8xm15*/
eiki7 = (MB & m15) && (C || !(MB & m13)); /* cin */
aleg = eastl? AR: 0; /* a input */
bleg = easbm? 0: DMASK; /* b input */
if (jamkn) /* jammin? xor */
D = aleg ^ bleg;
else D = (aleg + bleg + eiki7) & DMASK; /* else add */
/* Possible repeat at end of tlate - special t2, repeat tlate */
if (setaz) {
ARx = D; /* forced: t2 */
aleg = ARx; /* forced: tlate */
bleg = 0; /* forced */
jamkn = 0; /* forced */
D = (aleg + bleg + eiki7) & DMASK; /* forced add */
sc = 0; /* ends repeat */
}
/* Phase t4 */
clatr = (MB & (m11+m15+m16)) != 0; /* m11+m15+m16 */
cla1r = (MB & (m10+m14)) != 0; /* m10+m14 */
edahs = ((MB & (m11+m14)) == (m11+m14)) || /* (m11xm14)+m15+m16 */
(MB & (m15+m16));
edals = ((MB & (m11+m13)) == (m11+m13)) || /* (m11xm13)+m15+m16 */
(MB & (m15+m16));
etahs = (MB & (m9+m11)) == (m9+m11); /* m9xm11 */
etals = (MB & (m10+m11)) == (m10+m11); /* m10xm11 */
eda1r = ((MB & (m8+m10)) == (m8+m10)) || (MB & m14); /* (m8xm10)+m14 */
cbitl = (MB & (m9+m11)) == m9; /* m9x!m11 */
cbite = (MB & (m8+m9)) == (m8+m9); /* m8xm9 */
cbitg = (MB & (m10+m12)) == (m10+m12); /* m10xm12 */
if (clatr) /* clear A */
ARx = 0;
if (cla1r) /* clear A1 */
ARx = ARx & ~SIGN;
if (edahs) /* D hi to A hi */
ARx = ARx | (D & 0177400);
if (edals) /* D lo to A lo */
ARx = ARx | (D & 0000377);
if (etahs) /* D lo to A hi */
ARx = ARx | ((D << 8) & 0177400);
if (etals) /* D hi to A lo */
ARx = ARx | ((D >> 8) & 0000377);
if (eda1r) /* D1 to A1 */
ARx = ARx | (D & SIGN);
if (cbitl) { /* ovflo to C */
/* Overflow calculation. Cases:
aleg bleg cin overflow
0 x x can't overflow
A 0 0 can't overflow
A -1 1 can't overflow
A 0 1 overflow if 77777->100000
A -1 0 overflow if 100000->77777
*/
if (!jamkn &&
((bleg && !eiki7 && (D == 0077777)) ||
(!bleg && eiki7 && (D == 0100000))))
C = 1;
else C = 0;
}
if (cbite || (cbitg && (D & SIGN))) /* C = 1 */
C = 1;
return ARx;
}
/* Reset routines */
t_stat cpu_reset (DEVICE *dptr)
{
int32 i;
saved_AR = saved_BR = saved_XR = 0;
C = 0;
dp = 0;
ext = pme = extoff_pending = 0;
dev_int = dev_int & ~(INT_PEND|INT_NMI);
dev_ext_int = dev_enb = dev_ext_enb = 0;
for (i = 0; i < DMA_MAX; i++)
dma_ad[i] = dma_wc[i] = dma_eor[i] = 0;
chan_req = 0;
pcq_r = find_reg ("PCQ", NULL, dptr);
if (pcq_r)
pcq_r->qptr = 0;
else return SCPE_IERR;
// [RLA] We now have two break types - "E" (break on execution) and also "W"
// [RLA] (break on write)...
sim_brk_types = SWMASK('W') | SWMASK('E');
sim_brk_dflt = SWMASK ('E');
return SCPE_OK;
}
/* Memory examine */
t_stat cpu_ex (t_value *vptr, t_addr addr, UNIT *uptr, int32 sw)
{
if (addr >= MEMSIZE)
return SCPE_NXM;
if (vptr != NULL)
*vptr = M[addr] & DMASK;
return SCPE_OK;
}
/* Memory deposit */
t_stat cpu_dep (t_value val, t_addr addr, UNIT *uptr, int32 sw)
{
if (addr >= MEMSIZE)
return SCPE_NXM;
if (addr == 0)
saved_XR = val & DMASK;
M[addr] = val & DMASK;
return SCPE_OK;
}
/* Option processors */
t_stat cpu_set_noext (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
if (MEMSIZE > (NX_AMASK + 1))
return SCPE_ARG;
return SCPE_OK;
}
t_stat cpu_set_size (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
int32 mc = 0;
uint32 i;
if ((val <= 0) || (val > MAXMEMSIZE) || ((val & 07777) != 0) ||
(((cpu_unit.flags & UNIT_EXT) == 0) && (val > (NX_AMASK + 1))))
return SCPE_ARG;
for (i = val; i < MEMSIZE; i++)
mc = mc | M[i];
if ((mc != 0) && (!get_yn ("Really truncate memory [N]?", FALSE)))
return SCPE_OK;
MEMSIZE = val;
for (i = MEMSIZE; i < MAXMEMSIZE; i++)
M[i] = 0;
return SCPE_OK;
}
/* [RLA] Set/Show number of interrupts supported */
t_stat cpu_set_interrupts (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
uint32 newint; t_stat ret;
if (cptr == NULL) return SCPE_ARG;
newint = get_uint (cptr, 10, 49, &ret);
if (ret != SCPE_OK) return ret;
if ((newint != 0) && (newint != 16)) return SCPE_ARG;
ext_ints = newint;
return SCPE_OK;
}
t_stat cpu_show_interrupts (FILE *st, UNIT *uptr, int32 val, CONST void *desc)
{
if (ext_ints == 0)
fprintf(st,"standard interrupts");
else
fprintf(st,"extended interrupts = %d", ext_ints);
return SCPE_OK;
}
t_stat cpu_set_nchan (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
uint32 i, newmax;
t_stat r;
if (cptr == NULL)
return SCPE_ARG;
newmax = get_uint (cptr, 10, DMA_MAX, &r); /* get new max */
if ((r != SCPE_OK) || (newmax == dma_nch)) /* err or no chg? */
return r;
dma_nch = newmax; /* set new max */
for (i = newmax; i < DMA_MAX; i++) { /* reset chan */
dma_ad[i] = dma_wc[i] = dma_eor[i] = 0;
chan_req = chan_req & ~(1 << i);
}
return SCPE_OK;
}
/* Show DMA channels */
t_stat cpu_show_nchan (FILE *st, UNIT *uptr, int32 val, CONST void *desc)
{
if (dma_nch)
fprintf (st, "DMA channels = %d", dma_nch);
else fprintf (st, "no DMA");
return SCPE_OK;
}
/* Show channel state */
t_stat cpu_show_dma (FILE *st, UNIT *uptr, int32 val, CONST void *desc)
{
if ((val < 0) || (val >= DMA_MAX))
return SCPE_IERR;
fputs ((dma_ad[val] & DMA_IN)? "Input": "Output", st);
fprintf (st, ", addr = %06o, count = %06o, ", dma_ad[val] & X_AMASK, dma_wc[val]);
fprintf (st, "end of range %s\n", (dma_eor[val]? "set": "clear"));
return SCPE_OK;
}
/* Set I/O device to IOBUS / DMA channel / DMC channel */
t_stat io_set_iobus (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
DEVICE *dptr;
DIB *dibp;
if (val || cptr || (uptr == NULL))
return SCPE_IERR;
dptr = find_dev_from_unit (uptr);
if (dptr == NULL)
return SCPE_IERR;
dibp = (DIB *) dptr->ctxt;
if (dibp == NULL)
return SCPE_IERR;
dibp->chan = 0;
return SCPE_OK;
}
t_stat io_set_dma (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
DEVICE *dptr;
DIB *dibp;
uint32 newc;
t_stat r;
if ((cptr == NULL) || (uptr == NULL))
return SCPE_IERR;
dptr = find_dev_from_unit (uptr);
if (dptr == NULL)
return SCPE_IERR;
dibp = (DIB *) dptr->ctxt;
if (dibp == NULL)
return SCPE_IERR;
if (dma_nch == 0)
return SCPE_NOFNC;
newc = get_uint (cptr, 10, DMA_MAX, &r); /* get new */
if ((r != SCPE_OK) || (newc == 0) || (newc > dma_nch))
return SCPE_ARG;
dibp->chan = (newc - DMA_MIN) + DMA_V_DMA1 + 1; /* store */
return SCPE_OK;
}
t_stat io_set_dmc (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
DEVICE *dptr;
DIB *dibp;
uint32 newc;
t_stat r;
if ((cptr == NULL) || (uptr == NULL))
return SCPE_IERR;
dptr = find_dev_from_unit (uptr);
if (dptr == NULL)
return SCPE_IERR;
dibp = (DIB *) dptr->ctxt;
if (dibp == NULL)
return SCPE_IERR;
if (!(cpu_unit.flags & UNIT_DMC))
return SCPE_NOFNC;
newc = get_uint (cptr, 10, DMC_MAX, &r); /* get new */
if ((r != SCPE_OK) || (newc == 0))
return SCPE_ARG;
dibp->chan = (newc - DMC_MIN) + DMC_V_DMC1 + 1; /* store */
return SCPE_OK;
}
/* Show channel configuration */
t_stat io_show_chan (FILE *st, UNIT *uptr, int32 val, CONST void *desc)
{
DEVICE *dptr;
DIB *dibp;
if (uptr == NULL)
return SCPE_IERR;
dptr = find_dev_from_unit (uptr);
if (dptr == NULL)
return SCPE_IERR;
dibp = (DIB *) dptr->ctxt;
if (dibp == NULL)
return SCPE_IERR;
if (dibp->chan == 0)
fprintf (st, "IO bus");
else if (dibp->chan < (DMC_V_DMC1 + 1))
fprintf (st, "DMA channel %d", dibp->chan);
else fprintf (st, "DMC channel %d", dibp->chan - DMC_V_DMC1);
return SCPE_OK;
}
/* Set up I/O dispatch and channel maps */
// [RLA] Check for DMC conflicts (on both DMC channels!) ...
t_bool set_chanmap (DEVICE *dptr, DIB *dibp, uint32 dno, uint32 chan)
{
if ((chan < DMC_V_DMC1) && (chan >= dma_nch)) {
sim_printf ("%s configured for DMA channel %d\n", sim_dname (dptr), chan + 1);
return TRUE;
}
if ((chan >= DMC_V_DMC1) && !(cpu_unit.flags & UNIT_DMC)) {
sim_printf ("%s configured for DMC, option disabled\n", sim_dname (dptr));
return TRUE;
}
if (chan_map[chan]) { /* channel conflict? */
sim_printf ("%s DMA/DMC channel conflict, devno = %02o\n", sim_dname (dptr), dno);
return TRUE;
}
chan_map[chan] = dno; /* channel back map */
return FALSE;
}
t_bool devtab_init (void)
{
DEVICE *dptr;
DIB *dibp;
uint32 i, j, dno;
for (i = 0; i < DEV_MAX; i++)
iotab[i] = NULL;
for (i = 0; i < (DMA_MAX + DMC_MAX); i++)
chan_map[i] = 0;
for (i = 0; (dptr = sim_devices[i]); i++) { /* loop thru devices */
dibp = (DIB *) dptr->ctxt; /* get DIB */
if ((dibp == NULL) || (dptr->flags & DEV_DIS)) /* exist, enabled? */
continue;
dno = dibp->dev; /* device number */
for (j = 0; j < dibp->num; j++) { /* repeat for slots */
if (iotab[dno + j]) { /* conflict? */
sim_printf ("%s device number conflict, devno = %02o\n",
sim_dname (dptr), dno + j);
return TRUE;
}
iotab[dno + j] = dibp->io; /* set I/O routine */
} /* end for */
// [RLA] set up the channel map
if (dibp->chan != 0)
if (set_chanmap(dptr, dibp, dno, dibp->chan-1))
return TRUE;
if (dibp->chan2 != 0)
if (set_chanmap(dptr, dibp, dno, dibp->chan2-1))
return TRUE;
// [RLA] If the device uses extended interrupts, check that they're enabled.
if ((dibp->inum != INT_V_NONE) && (dibp->inum >= INT_V_EXTD) && (ext_ints == 0)) {
sim_printf ("%s uses extended interrupts but that option is disabled\n", sim_dname (dptr));
return TRUE;
}
} /* end for */
for (i = 0; i < DEV_MAX; i++) { /* fill in blanks */
if (iotab[i] == NULL)
iotab[i] = &undio;
}
return FALSE;
}
/* Set history */
t_stat cpu_set_hist (UNIT *uptr, int32 val, CONST char *cptr, void *desc)
{
int32 i, lnt;
t_stat r;
if (cptr == NULL) {
for (i = 0; i < hst_lnt; i++)
hst[i].pc = 0;
hst_p = 0;
return SCPE_OK;
}
lnt = (int32) get_uint (cptr, 10, HIST_MAX, &r);
if ((r != SCPE_OK) || (lnt && (lnt < HIST_MIN)))
return SCPE_ARG;
hst_p = 0;
if (hst_lnt) {
free (hst);
hst_lnt = 0;
hst = NULL;
}
if (lnt) {
hst = (InstHistory *) calloc (lnt, sizeof (InstHistory));
if (hst == NULL)
return SCPE_MEM;
hst_lnt = lnt;
}
return SCPE_OK;
}
/* Show history */
t_stat cpu_show_hist (FILE *st, UNIT *uptr, int32 val, CONST void *desc)
{
int32 cr, k, di, op, lnt;
const char *cptr = (const char *) desc;
t_value sim_eval;
t_stat r;
InstHistory *h;
static uint8 has_opnd[16] = {
0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1
};
if (hst_lnt == 0) /* enabled? */
return SCPE_NOFNC;
if (cptr) {
lnt = (int32) get_uint (cptr, 10, hst_lnt, &r);
if ((r != SCPE_OK) || (lnt == 0))
return SCPE_ARG;
}
else lnt = hst_lnt;
di = hst_p - lnt; /* work forward */
if (di < 0)
di = di + hst_lnt;
fprintf (st, " PC C A B X ea IR\n");
fprintf (st, "----- - ------ ------ ------ ----- -----------\n\n");
for (k = 0; k < lnt; k++) { /* print specified */
h = &hst[(++di) % hst_lnt]; /* entry pointer */
if (h->pc & HIST_PC) { /* instruction? */
cr = (h->pc & HIST_C)? 1: 0; /* carry */
fprintf (st, "%05o %o %06o %06o %06o ",
h->pc & X_AMASK, cr, h->ar, h->br, h->xr);
if (h->pc & HIST_EA)
fprintf (st, "%05o ", h->ea);
else fprintf (st, " ");
sim_eval = h->ir;
if ((fprint_sym (st, h->pc & X_AMASK, &sim_eval,
&cpu_unit, SWMASK ('M'))) > 0)
fprintf (st, "(undefined) %06o", h->ir);
op = I_GETOP (h->ir) & 017; /* base op */
if (has_opnd[op])
fprintf (st, " [%06o]", h->opnd);
if (h->iack) // [RLA]
fprintf(st, " INTERRUPT"); // [RLA]
fputc ('\n', st); /* end line */
} /* end else instruction */
} /* end for */
return SCPE_OK;
}