Mercurial > hg-old > index.cgi
annotate lwasm/parse.c @ 190:563adfccb645
Added 'sym=expr' opcode handling
author | lost |
---|---|
date | Sun, 22 Mar 2009 16:08:20 +0000 |
parents | 99300be2d3bd |
children | bae1e3ecdce1 |
rev | line source |
---|---|
36 | 1 /* |
2 parse.c | |
3 Copyright © 2008 William Astle | |
4 | |
5 This file is part of LWASM. | |
6 | |
7 LWASM is free software: you can redistribute it and/or modify it under the | |
8 terms of the GNU General Public License as published by the Free Software | |
9 Foundation, either version 3 of the License, or (at your option) any later | |
10 version. | |
11 | |
12 This program is distributed in the hope that it will be useful, but WITHOUT | |
13 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for | |
15 more details. | |
16 | |
17 You should have received a copy of the GNU General Public License along with | |
18 this program. If not, see <http://www.gnu.org/licenses/>. | |
19 */ | |
20 | |
21 /* | |
22 Contains the general parser | |
23 */ | |
24 | |
25 #define __parse_c_seen__ | |
26 | |
27 #include <ctype.h> | |
28 #include <string.h> | |
29 | |
30 #include "lwasm.h" | |
31 #include "instab.h" | |
32 #include "util.h" | |
33 | |
34 // parse a line and despatch to the appropriate handlers for opcodes | |
35 int lwasm_parse_line(asmstate_t *as, lwasm_line_t *l) | |
36 { | |
37 char *p, *p2; | |
38 char *opc; | |
39 int opnum; | |
44 | 40 char *sym = NULL; |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
41 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
42 // if this was a bad op first pass (or otherwise a no-process line) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
43 // ignore it |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
44 if (l -> badop) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
45 return; |
36 | 46 |
47 p = l -> text; | |
101
f59c0916753d
Fixed relative branches and PCR addressing to handle constant intra-section references properly
lost
parents:
99
diff
changeset
|
48 l -> sect = as -> csect; |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
49 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
50 // blank lines are a no brainer |
36 | 51 if (!*p) |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
52 { |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
53 as -> context = lwasm_next_context(as); |
36 | 54 return 0; |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
55 } |
36 | 56 |
52 | 57 // for output generation later but only on pass 1 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
58 // also used by some pseudo ops on pass 2 |
52 | 59 if (as -> passnum == 1) |
60 l -> codeaddr = as -> addr; | |
44 | 61 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
62 // if it's a comment, return (this doesn't cause a context change) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
63 if (*p == '*' || *p == ';') |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
64 return; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
65 |
169 | 66 // if we have C pre-processor directives/output, ignore it |
67 if (*p == '#') | |
68 return; | |
69 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
70 // if we start with a non-space character, it's a symbol |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
71 if (!isspace(*p)) |
36 | 72 { |
73 // we have a symbol specified here | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
74 // parse it out and record it for later use |
157 | 75 for (p2 = p; *p2 && !isspace(*p2) && *p2 != ':'; p2++) |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
76 /* do nothing */ ; |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
77 |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
78 sym = lwasm_alloc((p2 - p) + 1); |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
79 sym[p2 - p] = '\0'; |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
80 memcpy(sym, p, p2 - p); |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
81 |
44 | 82 p = p2; |
157 | 83 if (!*sym) |
84 { | |
85 register_error(as, l, 1, "Invalid symbol"); | |
86 lwasm_free(sym); | |
87 sym = NULL; | |
88 } | |
89 if (*p == ':') | |
90 p++; | |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
91 } |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
92 l -> sym = sym; |
36 | 93 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
94 // now skip any whitespace to find the opcode |
36 | 95 while (*p && isspace(*p)) |
96 p++; | |
97 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
98 // is the line blank? |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
99 if (!*p && !sym) |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
100 { |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
101 // nothing but white space *is* a context break |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
102 as -> context = lwasm_next_context(as); |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
103 return; |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
104 } |
36 | 105 |
106 // parse the opcode | |
107 for (p2 = p; *p2 && !isspace(*p2); p2++) | |
108 /* do nothing */ ; | |
109 | |
110 opc = lwasm_alloc((p2 - p) + 1); | |
111 memcpy(opc, p, p2 - p); | |
112 opc[p2 - p] = '\0'; | |
113 | |
190 | 114 l -> forceglobal = 0; |
115 // if the opcode contains an =, treat it as "symbol = expr" | |
116 if (!sym && strchr(opc, '=')) | |
117 { | |
118 for (p2 = opc; *p2 && *p2 != '='; p2++) | |
119 /* do nothing */ ; | |
120 sym = lwasm_alloc((p2 - opc) + 1); | |
121 memcpy(sym, opc, p2 - opc); | |
122 sym[p2 - opc] = '\0'; | |
123 l -> sym = sym; | |
124 | |
125 p2 = p + (p2 - opc) + 1; | |
126 // p2++; | |
127 opc[0] = '='; | |
128 opc[1] = '\0'; | |
129 debug_message(2, "Found opcode = with symbol %s and operand %s", sym, p2); | |
130 l -> forceglobal = 1; | |
131 } | |
132 | |
38 | 133 debug_message(2, "Found operation code: '%s'", opc); |
134 | |
36 | 135 // skip intervening whitespace if present |
136 while (*p2 && isspace(*p2)) | |
137 p2++; | |
138 | |
139 // look up instruction in insn table | |
140 for (opnum = 0; instab[opnum].opcode; opnum++) | |
141 { | |
142 if (!strcasecmp(instab[opnum].opcode, opc)) | |
143 break; | |
144 } | |
145 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
146 // if we found no operation, check if we had a comment |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
147 // the reason this check is here is to allow for "private" |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
148 // operation codes like "*pragma" which will be ignored by |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
149 // other assemblers |
68 | 150 // also skip empty ops |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
151 if (!(instab[opnum].opcode)) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
152 { |
68 | 153 if (*opc == '*' || *opc == ';' || !*opc) |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
154 goto done_line; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
155 } |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
156 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
157 // now we have the opcode and the symbol, we can decide if we're |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
158 // actually going to do anything with this line |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
159 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
160 // we will NOT call the function if any of the following are true: |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
161 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
162 // - we are skipping a condition and the operation code is not a conditional |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
163 // - we are defining a macro and the operation code is not ENDM |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
164 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
165 // we will call the function in any other circumstance |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
166 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
167 // first condition above |
57 | 168 if (as -> inmacro && instab[opnum].endm == 0) |
169 { | |
170 add_macro_line(as, l -> text); | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
171 goto done_line; |
57 | 172 } |
173 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
174 // second condition above |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
175 if (as -> skipcond && instab[opnum].iscond == 0) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
176 goto done_line; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
177 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
178 // we've registered the symbol as needed |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
179 // now we need to check for a macro call IFF we don't collide with |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
180 // an operation code; otherwise, call the operation function |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
181 if (instab[opnum].opcode) |
36 | 182 { |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
183 if (instab[opnum].fn) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
184 { |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
185 (instab[opnum].fn)(as, l, &p2, opnum); |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
186 } |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
187 else |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
188 { |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
189 // carp about unimplemented operation |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
190 register_error(as, l, 1, "Unimplemented operation code: %s", opc); |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
191 } |
36 | 192 } |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
193 else |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
194 { |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
195 if (expand_macro(as, l, &p2, opc) == 0) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
196 goto done_line; |
36 | 197 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
198 // carp about an unknown operation code and note that fact for |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
199 // pass 2 in case a macro appears later with the same name! |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
200 register_error(as, l, 1, "Uknown operation code: %s", opc); |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
201 l -> badop = 1; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
202 } |
57 | 203 |
204 done_line: | |
68 | 205 if (!(as -> skipcond || as -> inmacro)) |
206 { | |
207 // register symbol if the operation didn't | |
208 if (sym && instab[opnum].setsym == 0) | |
209 { | |
210 if (as -> passnum == 1) | |
211 { | |
71
90a5657d5408
Fixed problem with symbols being registered with the address of the NEXT instruction
lost
parents:
68
diff
changeset
|
212 debug_message(1, "Registering symbol '%s' at %04X", sym, l -> codeaddr); |
90a5657d5408
Fixed problem with symbols being registered with the address of the NEXT instruction
lost
parents:
68
diff
changeset
|
213 if (lwasm_register_symbol(as, l, sym, l -> codeaddr, SYMBOL_NORM) < 0) |
68 | 214 l -> sym = NULL; |
215 else | |
216 l -> addrset = 1; | |
217 } | |
218 } | |
219 } | |
99 | 220 |
85 | 221 l -> sect = as -> csect; |
91
718998b673ee
Added incomplete references to object output and added support for section base terms in expression handler
lost
parents:
85
diff
changeset
|
222 l -> context = as -> context; |
85 | 223 |
36 | 224 lwasm_free(opc); |
44 | 225 if (sym) |
226 lwasm_free(sym); | |
36 | 227 } |