freebsd-dev/sys/dev/aic7xxx/aicasm.c

686 lines
15 KiB
C
Raw Normal View History

/*+M*************************************************************************
* Adaptec AIC7770/AIC7870 sequencer code assembler.
*
* Copyright (c) 1994 John Aycock
* The University of Calgary Department of Computer Science.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions, and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of Calgary
* Department of Computer Science and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* Comments are started by `#' and continue to the end of the line; lines
* may be of the form:
* <label>*
* <label>* <undef-sym> = <value>
* <label>* <opcode> <operand>*
*
* A <label> is an <undef-sym> ending in a colon. Spaces, tabs, and commas
* are token separators.
*
*-M*************************************************************************/
static char id[] = "$Id: aic7xxx_asm.c,v 1.11 1996/01/05 01:48:07 gibbs Exp $";
#include <ctype.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
#include <fcntl.h>
#define MEMORY 448
#define MAXLINE 1024
#define MAXTOKEN 32
#define ADOTOUT "a.out"
#define NOVALUE -1
/*
* AIC-7770/AIC-7870 register definitions
*/
#define R_SINDEX 0x65
#define R_ALLONES 0x69
#define R_ALLZEROS 0x6a
#define R_NONE 0x6a
int debug;
int lineno, LC;
char *filename;
unsigned char M[MEMORY][4];
1995-05-30 08:16:23 +00:00
void
error(char *s)
{
fprintf(stderr, "%s: %s at line %d\n", filename, s, lineno);
exit(EXIT_FAILURE);
}
void *
Malloc(size_t size)
{
void *p = malloc(size);
if (!p)
error("out of memory");
return(p);
}
void *
Realloc(void *ptr, size_t size)
{
void *p = realloc(ptr, size);
if (!p)
error("out of memory");
return(p);
}
char *
Strdup(char *s)
{
char *p = (char *)Malloc(strlen(s) + 1);
strcpy(p, s);
return(p);
}
typedef struct sym_t {
struct sym_t *next; /* MUST BE FIRST */
char *name;
int value;
1995-05-30 08:16:23 +00:00
int npatch;
int *patch;
} sym_t;
sym_t *head;
void
define(char *name, int value)
{
sym_t *p, *q;
for (p = head, q = (sym_t *)&head; p; p = p->next) {
if (!strcmp(p->name, name))
error("redefined symbol");
q = p;
}
p = q->next = (sym_t *)Malloc(sizeof(sym_t));
p->next = NULL;
p->name = Strdup(name);
p->value = value;
p->npatch = 0;
p->patch = NULL;
if (debug) {
fprintf(stderr, "\"%s\" ", p->name);
if (p->value != NOVALUE)
fprintf(stderr, "defined as 0x%x\n", p->value);
else
fprintf(stderr, "undefined\n");
}
}
sym_t *
lookup(char *name)
{
sym_t *p;
for (p = head; p; p = p->next)
if (!strcmp(p->name, name))
return(p);
return(NULL);
}
1995-05-30 08:16:23 +00:00
void
patch(sym_t *p, int location)
{
p->npatch += 1;
p->patch = (int *)Realloc(p->patch, p->npatch * sizeof(int *));
p->patch[p->npatch - 1] = location;
}
void backpatch(void)
{
int i;
sym_t *p;
for (p = head; p; p = p->next) {
if (p->value == NOVALUE) {
fprintf(stderr,
"%s: undefined symbol \"%s\"\n",
filename, p->name);
exit(EXIT_FAILURE);
}
if (p->npatch) {
if (debug)
fprintf(stderr,
"\"%s\" (0x%x) patched at",
p->name, p->value);
for (i = 0; i < p->npatch; i++) {
M[p->patch[i]][0] &= ~1;
M[p->patch[i]][0] |= ((p->value >> 8) & 1);
M[p->patch[i]][1] = p->value & 0xff;
if (debug)
fprintf(stderr, " 0x%x", p->patch[i]);
}
if (debug)
fputc('\n', stderr);
}
}
}
/*
* Output words in byte-reversed order (least significant first)
* since the sequencer RAM is loaded that way.
*/
void
output(FILE *fp)
{
int i;
for (i = 0; i < LC; i++)
fprintf(fp, "\t0x%02x, 0x%02x, 0x%02x, 0x%02x,\n",
M[i][3],
M[i][2],
M[i][1],
M[i][0]);
printf("%d out of %d instructions used.\n", LC, MEMORY);
}
char **
getl(int *n)
{
int i;
char *p, *quote;
static char buf[MAXLINE];
static char *a[MAXTOKEN];
i = 0;
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
while (fgets(buf, sizeof(buf), stdin)) {
lineno += 1;
if (buf[strlen(buf)-1] != '\n')
error("line too long");
p = strchr(buf, '#');
if (p)
*p = '\0';
p = buf;
rescan:
quote = strchr(p, '\"');
if (quote)
*quote = '\0';
for (p = strtok(p, ", \t\n"); p; p = strtok(NULL, ", \t\n"))
if (i < MAXTOKEN-1)
a[i++] = p;
else
error("too many tokens");
if (quote) {
1995-05-30 08:16:23 +00:00
quote++;
p = strchr(quote, '\"');
if (!p)
error("unterminated string constant");
else if (i < MAXTOKEN-1) {
a[i++] = quote;
*p = '\0';
p++;
}
else
error("too many tokens");
goto rescan;
1995-05-30 08:16:23 +00:00
}
if (i) {
*n = i;
return(a);
}
}
return(NULL);
}
#define A 0x8000 /* `A'ccumulator ok */
#define I 0x4000 /* use as immediate value */
#define SL 0x2000 /* shift left */
#define SR 0x1000 /* shift right */
#define RL 0x0800 /* rotate left */
#define RR 0x0400 /* rotate right */
#define LO 0x8000 /* lookup: ori-{jmp,jc,jnc,call} */
#define LA 0x4000 /* lookup: and-{jz,jnz} */
#define LX 0x2000 /* lookup: xor-{je,jne} */
#define NA -1 /* not applicable */
struct {
char *name;
int n; /* number of operands, including opcode */
unsigned int op; /* immediate or L?|pos_from_0 */
unsigned int dest; /* NA, pos_from_0, or I|immediate */
unsigned int src; /* NA, pos_from_0, or I|immediate */
unsigned int imm; /* pos_from_0, A|pos_from_0, or I|immediate */
unsigned int addr; /* NA or pos_from_0 */
int fmt; /* instruction format - 1, 2, or 3 */
} instr[] = {
/*
* N OP DEST SRC IMM ADDR FMT
*/
{ "mov", 3, 1, 1, 2, I|0xff, NA, 1 },
{ "mov", 4, LO|2, NA, 1, I|0, 3, 3 },
{ "mvi", 3, 0, 1, I|R_ALLZEROS, A|2, NA, 1 },
{ "mvi", 4, LO|2, NA, I|R_ALLZEROS, 1, 3, 3 },
{ "not", 2, 2, 1, 1, I|0xff, NA, 1 },
{ "and", 3, 1, 1, 1, A|2, NA, 1 },
{ "and", 4, 1, 1, 3, A|2, NA, 1 },
{ "or", 3, 0, 1, 1, A|2, NA, 1 },
{ "or", 4, 0, 1, 3, A|2, NA, 1 },
{ "or", 5, LO|3, NA, 1, 2, 4, 3 },
{ "xor", 3, 2, 1, 1, A|2, NA, 1 },
{ "xor", 4, 2, 1, 3, A|2, NA, 1 },
{ "nop", 1, 1, I|R_NONE, I|R_ALLZEROS, I|0xff, NA, 1 },
{ "inc", 2, 3, 1, 1, I|1, NA, 1 },
{ "inc", 3, 3, 1, 2, I|1, NA, 1 },
{ "dec", 2, 3, 1, 1, I|0xff, NA, 1 },
{ "dec", 3, 3, 1, 2, I|0xff, NA, 1 },
{ "jmp", 2, LO|0, NA, I|R_SINDEX, I|0, 1, 3 },
{ "jc", 2, LO|0, NA, I|R_SINDEX, I|0, 1, 3 },
{ "jnc", 2, LO|0, NA, I|R_SINDEX, I|0, 1, 3 },
{ "call", 2, LO|0, NA, I|R_SINDEX, I|0, 1, 3 },
{ "test", 5, LA|3, NA, 1, A|2, 4, 3 },
{ "cmp", 5, LX|3, NA, 1, A|2, 4, 3 },
{ "ret", 1, 1, I|R_NONE, I|R_ALLZEROS, I|0xff, NA, 1 },
{ "ret", 1, 1, I|R_NONE, I|R_ALLZEROS, I|0xff, NA, 1 },
{ "clc", 1, 3, I|R_NONE, I|R_ALLZEROS, I|1, NA, 1 },
{ "clc", 4, 3, 2, I|R_ALLZEROS, A|3, NA, 1 },
{ "stc", 2, 3, 1, I|R_ALLONES, I|1, NA, 1 },
{ "add", 3, 3, 1, 1, A|2, NA, 1 },
{ "add", 4, 3, 1, 3, A|2, NA, 1 },
{ "adc", 3, 4, 1, 1, A|2, NA, 1 },
{ "adc", 4, 4, 1, 3, A|2, NA, 1 },
{ "shl", 3, 5, 1, 1, SL|2, NA, 2 },
{ "shl", 4, 5, 1, 2, SL|3, NA, 2 },
{ "shr", 3, 5, 1, 1, SR|2, NA, 2 },
{ "shr", 4, 5, 1, 2, SR|3, NA, 2 },
{ "rol", 3, 5, 1, 1, RL|2, NA, 2 },
{ "rol", 4, 5, 1, 2, RL|3, NA, 2 },
{ "ror", 3, 5, 1, 1, RR|2, NA, 2 },
{ "ror", 4, 5, 1, 2, RR|3, NA, 2 },
/*
* Extensions (note also that mvi allows A)
*/
{ "clr", 2, 1, 1, I|R_ALLZEROS, I|0xff, NA, 1 },
{ 0, 0, 0, 0, 0, 0, 0, 0 }
};
1995-05-30 08:16:23 +00:00
int
eval_operand(char **a, int spec)
{
int i;
unsigned int want = spec & (LO|LA|LX);
static struct {
unsigned int what;
char *name;
int value;
} jmptab[] = {
{ LO, "jmp", 8 },
{ LO, "jc", 9 },
{ LO, "jnc", 10 },
{ LO, "call", 11 },
{ LA, "jz", 15 },
{ LA, "jnz", 13 },
{ LX, "je", 14 },
{ LX, "jne", 12 },
};
spec &= ~(LO|LA|LX);
for (i = 0; i < sizeof(jmptab)/sizeof(jmptab[0]); i++)
if (jmptab[i].what == want &&
!strcmp(jmptab[i].name, a[spec]))
{
return(jmptab[i].value);
}
if (want)
error("invalid jump");
return(spec); /* "case 0" - no flags set */
}
int
eval_sdi(char **a, int spec)
{
sym_t *p;
unsigned val;
if (spec == NA)
return(NA);
switch (spec & (A|I|SL|SR|RL|RR)) {
case SL:
case SR:
case RL:
case RR:
if (isdigit(*a[spec &~ (SL|SR|RL|RR)]))
val = strtol(a[spec &~ (SL|SR|RL|RR)], NULL, 0);
else {
p = lookup(a[spec &~ (SL|SR|RL|RR)]);
if (!p)
error("undefined symbol used");
val = p->value;
}
switch (spec & (SL|SR|RL|RR)) { /* blech */
case SL:
if (val > 7)
return(0xf0);
return(((val % 8) << 4) |
(val % 8));
case SR:
if (val > 7)
return(0xf0);
return(((val % 8) << 4) |
(1 << 3) |
((8 - (val % 8)) % 8));
case RL:
return(val % 8);
case RR:
return((8 - (val % 8)) % 8);
}
case I:
return(spec &~ I);
case A:
/*
* An immediate field of zero selects
* the accumulator. Vigorously object
* if zero is given otherwise - it's
* most likely an error.
*/
spec &= ~A;
if (!strcmp("A", a[spec]))
return(0);
if (isdigit(*a[spec]) &&
strtol(a[spec], NULL, 0) == 0)
{
error("immediate value of zero selects accumulator");
}
/* falls through */
case 0:
if (isdigit(*a[spec]))
return(strtol(a[spec], NULL, 0));
p = lookup(a[spec]);
if (p)
return(p->value);
error("undefined symbol used");
}
return(NA); /* shut the compiler up */
}
int
eval_addr(char **a, int spec)
{
sym_t *p;
if (spec == NA)
return(NA);
if (isdigit(*a[spec]))
return(strtol(a[spec], NULL, 0));
p = lookup(a[spec]);
if (p) {
if (p->value != NOVALUE)
return(p->value);
patch(p, LC);
} else {
define(a[spec], NOVALUE);
p = lookup(a[spec]);
patch(p, LC);
}
return(NA); /* will be patched in later */
}
int
crack(char **a, int n)
{
int i;
int I_imm, I_addr;
int I_op, I_dest, I_src, I_ret;
/*
* Check for "ret" at the end of the line; remove
* it unless it's "ret" alone - we still want to
* look it up in the table.
*/
I_ret = (strcmp(a[n-1], "ret") ? 0 : !0);
if (I_ret && n > 1)
n -= 1;
for (i = 0; instr[i].name; i++) {
/*
* Look for match in table given constraints,
* currently just the name and the number of
* operands.
*/
if (!strcmp(instr[i].name, *a) && instr[i].n == n)
break;
}
if (!instr[i].name)
error("unknown opcode or wrong number of operands");
I_op = eval_operand(a, instr[i].op);
I_src = eval_sdi(a, instr[i].src);
I_imm = eval_sdi(a, instr[i].imm);
I_dest = eval_sdi(a, instr[i].dest);
I_addr = eval_addr(a, instr[i].addr);
if( LC >= MEMORY )
error("Memory exhausted!\n");
switch (instr[i].fmt) {
case 1:
case 2:
M[LC][0] = (I_op << 1) | I_ret;
M[LC][1] = I_dest;
M[LC][2] = I_src;
M[LC][3] = I_imm;
break;
case 3:
if (I_ret)
error("illegal use of \"ret\"");
M[LC][0] = (I_op << 1) | ((I_addr >> 8) & 1);
M[LC][1] = I_addr & 0xff;
M[LC][2] = I_src;
M[LC][3] = I_imm;
break;
}
return (1); /* no two-byte instructions yet */
}
#undef SL
#undef SR
#undef RL
#undef RR
#undef LX
#undef LA
#undef LO
#undef I
#undef A
void
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
assemble(FILE *ofile)
{
int n;
char **a;
sym_t *p;
while ((a = getl(&n))) {
while (a[0][strlen(*a)-1] == ':') {
a[0][strlen(*a)-1] = '\0';
p = lookup(*a);
if (p)
p->value = LC;
else
define(*a, LC);
a += 1;
n -= 1;
}
if (!n) /* line was all labels */
continue;
if (n == 3 && !strcmp("VERSION", *a))
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
fprintf(ofile, "#define %s \"%s\"\n", a[1], a[2]);
else {
if (n == 3 && !strcmp("=", a[1]))
define(*a, strtol(a[2], NULL, 0));
else
LC += crack(a, n);
}
}
backpatch();
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
output(ofile);
if (debug)
output(stderr);
}
int
main(int argc, char **argv)
{
int c;
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
int pid;
int ifile;
FILE *ofile;
int fd[2];
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
ofile = NULL;
while ((c = getopt(argc, argv, "dho:vD:")) != EOF) {
switch (c) {
case 'd':
debug = !0;
break;
case 'D':
{
char *p;
if ((p = strchr(optarg, '=')) != NULL) {
*p = '\0';
define(optarg, strtol(p + 1, NULL, 0));
}
else
define(optarg, 1);
break;
}
case 'o':
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
if ((ofile = fopen(optarg, "w")) == NULL) {
perror(optarg);
exit(EXIT_FAILURE);
}
break;
case 'h':
1995-05-30 08:16:23 +00:00
printf("usage: %s [-d] [-Dname] [-ooutput] input\n",
*argv);
exit(EXIT_SUCCESS);
break;
case 'v':
printf("%s\n", id);
exit(EXIT_SUCCESS);
break;
default:
exit(EXIT_FAILURE);
break;
}
}
if (argc - optind != 1) {
fprintf(stderr, "%s: must have one input file\n", *argv);
exit(EXIT_FAILURE);
}
filename = argv[optind];
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
if ((ifile = open(filename, O_RDONLY)) < 0) {
perror(filename);
exit(EXIT_FAILURE);
}
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
if (!ofile) {
if ((ofile = fopen(ADOTOUT, "w")) == NULL) {
perror(ADOTOUT);
exit(EXIT_FAILURE);
}
}
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
if (pipe(fd) < 0) {
perror("pipe failed");
exit(1);
}
if ((pid = fork()) < 0 ) {
perror("fork failed");
exit(1);
}
else if (pid > 0) { /* Parent */
close(fd[1]); /* Close write end */
if (fd[0] != STDIN_FILENO) {
if (dup2(fd[0], STDIN_FILENO) != STDIN_FILENO) {
perror("dup2 error on stdin");
exit(EXIT_FAILURE);
}
close(fd[0]);
}
assemble(ofile);
exit(EXIT_SUCCESS);
}
else { /* Child */
close(fd[0]); /* Close Read end */
if (fd[1] != STDOUT_FILENO) {
if (dup2(fd[1], STDOUT_FILENO) != STDOUT_FILENO) {
perror("dup2 error on stdout");
exit(EXIT_FAILURE);
}
close(fd[1]);
}
if (ifile != STDIN_FILENO) {
if (dup2(ifile, STDIN_FILENO) != STDIN_FILENO) {
perror("dup2 error on stdin");
exit(EXIT_FAILURE);
}
close(ifile);
}
execl("/usr/bin/cpp", "/usr/bin/cpp", "-P", "-", "-", NULL);
The long awaited stability patch set for the aic7xxx driver: 1) Use cpp to preprocess the sequencer code. 2) Convert all "magic numbers" to #defines shared by the sequencer and kernel driver via the aic7xxx_reg.h file. (The assembler still needs to be re-written in lex/yacc to allow ~|& type constructions). 3) Raise ATN on parity errors for "in" phases and send an initiator detected error or message-in parity error message as appropriate. 4) Turn off the reselection hardware from the time or a (re)connection to busfree. It seems that some fast targets were able to reconnect before the sequencer was able to see busfree. 5) The message buffer is considered "in-use" when there is a positive length count. The ACTIVE_MSG flag was unnecesary. 6) Properly set SCB_NEXT_WAITING to SCB_LIST_HEAD in scbs being added to the waiting scb list. This is a change in how the list code works to facilitate some planned work in the reset code. 7) The fields in the SCB have be re-arranged to be quad-word aligned. 8) The inb code has been rewritten to catch phasemisses and be more efficient. 9) Go back to "snooping the bus" to determine if the incomming identify message will be followed by a simple queue message. Its much faster than doing a search through the SCBs. 10) Implement better tag range checking for incomming tags. 11) Make sdtr_to_rate more accurate (use 25 instead of 24 in calculations - must have been asleep that night). 12) Rearrange some routines to reduce code complexity and size. 13) Update comments and formatting. 14) Fixed bugs I've forgotten about?? Reviewed by: David Greenman <davidg@FreeBSD.org>
1996-01-03 06:25:32 +00:00
}
return(EXIT_SUCCESS);
}