|
116
|
1 /*
|
|
|
2 readfiles.c
|
|
118
|
3 Copyright © 2009 William Astle
|
|
116
|
4
|
|
|
5 This file is part of LWLINK.
|
|
|
6
|
|
|
7 LWLINK 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 Reads input files
|
|
|
22
|
|
|
23 */
|
|
|
24
|
|
|
25 #ifdef HAVE_CONFIG_H
|
|
|
26 #include "config.h"
|
|
|
27 #endif
|
|
|
28
|
|
|
29 #include <argp.h>
|
|
|
30 #include <errno.h>
|
|
|
31 #include <stdio.h>
|
|
|
32 #include <stdlib.h>
|
|
|
33
|
|
|
34 #include "lwlink.h"
|
|
|
35 #include "util.h"
|
|
|
36
|
|
|
37 void read_lwobj16v0(fileinfo_t *fn);
|
|
|
38
|
|
|
39 /*
|
|
|
40 The logic of reading the entire file into memory is simple. All the symbol
|
|
|
41 names in the file are NUL terminated strings and can be used directly without
|
|
|
42 making additional copies.
|
|
|
43 */
|
|
|
44 void read_files(void)
|
|
|
45 {
|
|
|
46 int i;
|
|
|
47 long size;
|
|
|
48 FILE *f;
|
|
|
49 long bread;
|
|
|
50 for (i = 0; i < ninputfiles; i++)
|
|
|
51 {
|
|
|
52 f = fopen(inputfiles[i] -> filename, "rb");
|
|
|
53 if (!f)
|
|
|
54 {
|
|
|
55 fprintf(stderr, "Can't open file %s:", inputfiles[i] -> filename);
|
|
|
56 perror("");
|
|
|
57 exit(1);
|
|
|
58 }
|
|
|
59 fseek(f, 0, SEEK_END);
|
|
|
60 size = ftell(f);
|
|
|
61 rewind(f);
|
|
|
62
|
|
|
63 inputfiles[i] -> filedata = lw_malloc(size);
|
|
|
64 inputfiles[i] -> filesize = size;
|
|
|
65
|
|
|
66 bread = fread(inputfiles[i] -> filedata, 1, size, f);
|
|
|
67 if (bread < size)
|
|
|
68 {
|
|
|
69 fprintf(stderr, "Short read on file %s (%ld/%ld):", inputfiles[i] -> filename, bread, size);
|
|
|
70 perror("");
|
|
|
71 exit(1);
|
|
|
72 }
|
|
|
73
|
|
|
74 fclose(f);
|
|
|
75
|
|
|
76 if (!memcmp(inputfiles[i] -> filedata, "LWOBJ16", 8))
|
|
|
77 {
|
|
|
78 // read v0 LWOBJ16 file
|
|
|
79 read_lwobj16v0(inputfiles[i]);
|
|
|
80 }
|
|
|
81 else
|
|
|
82 {
|
|
|
83 fprintf(stderr, "%s: unknown file format\n", inputfiles[i] -> filename);
|
|
|
84 exit(1);
|
|
|
85 }
|
|
|
86 }
|
|
|
87 }
|
|
|
88
|
|
|
89 // this macro is used to bail out if we run off the end of the file data
|
|
|
90 // while parsing - it keeps the code below cleaner
|
|
124
|
91 #define NEXTBYTE() do { cc++; if (cc > fn -> filesize) { fprintf(stderr, "%s: invalid file format\n", fn -> filename); exit(1); } } while (0)
|
|
116
|
92 // this macro is used to refer to the current byte in the stream
|
|
124
|
93 #define CURBYTE() (fn -> filedata[cc < fn -> filesize ? cc : fn -> filesize - 1])
|
|
116
|
94 // this one will leave the input pointer past the trailing NUL
|
|
|
95 #define CURSTR() read_lwobj16v0_str(&cc, fn)
|
|
|
96 unsigned char *read_lwobj16v0_str(long *cc1, fileinfo_t *fn)
|
|
|
97 {
|
|
|
98 int cc = *cc1;
|
|
|
99 unsigned char *fp;
|
|
|
100 fp = &CURBYTE();
|
|
|
101 while (CURBYTE())
|
|
|
102 NEXTBYTE();
|
|
|
103 NEXTBYTE();
|
|
|
104 *cc1 = cc;
|
|
|
105 return fp;
|
|
|
106 }
|
|
|
107 // the function below can be switched to dealing with data coming from a
|
|
|
108 // source other than an in-memory byte pool by adjusting the input data
|
|
|
109 // in "fn" and the above two macros
|
|
|
110 void read_lwobj16v0(fileinfo_t *fn)
|
|
|
111 {
|
|
|
112 unsigned char *fp;
|
|
|
113 long cc;
|
|
|
114 section_t *s;
|
|
|
115 int val;
|
|
|
116 symtab_t *se;
|
|
|
117
|
|
|
118 // start reading *after* the magic number
|
|
126
|
119 cc = 8;
|
|
116
|
120
|
|
|
121 // init data
|
|
|
122 fn -> sections = NULL;
|
|
|
123 fn -> nsections = 0;
|
|
|
124
|
|
|
125 while (1)
|
|
|
126 {
|
|
126
|
127 // NEXTBYTE();
|
|
116
|
128 // bail out if no more sections
|
|
|
129 if (!(CURBYTE()))
|
|
|
130 break;
|
|
|
131
|
|
|
132 fp = CURSTR();
|
|
|
133
|
|
|
134 // we now have a section name in fp
|
|
|
135 // create new section entry
|
|
|
136 fn -> sections = lw_realloc(fn -> sections, sizeof(section_t) * (fn -> nsections + 1));
|
|
|
137 s = &(fn -> sections[fn -> nsections]);
|
|
|
138 fn -> nsections += 1;
|
|
|
139
|
|
|
140 s -> localsyms = NULL;
|
|
|
141 s -> flags = 0;
|
|
|
142 s -> codesize = 0;
|
|
|
143 s -> name = fp;
|
|
|
144 s -> loadaddress = 0;
|
|
|
145 s -> localsyms = NULL;
|
|
|
146 s -> exportedsyms = NULL;
|
|
|
147 s -> incompletes = NULL;
|
|
119
|
148 s -> processed = 0;
|
|
|
149 s -> file = fn;
|
|
|
150
|
|
116
|
151 // read flags
|
|
|
152 while (CURBYTE())
|
|
|
153 {
|
|
|
154 switch (CURBYTE())
|
|
|
155 {
|
|
|
156 case 0x01:
|
|
|
157 s -> flags |= SECTION_BSS;
|
|
|
158 break;
|
|
|
159
|
|
|
160 default:
|
|
|
161 fprintf(stderr, "%s (%s): unrecognized section flag %02X\n", fn -> filename, s -> name, (int)(CURBYTE()));
|
|
|
162 exit(1);
|
|
|
163 }
|
|
|
164 NEXTBYTE();
|
|
|
165 }
|
|
|
166 // skip NUL terminating flags
|
|
|
167 NEXTBYTE();
|
|
|
168
|
|
|
169 // now parse the local symbol table
|
|
|
170 while (CURBYTE())
|
|
|
171 {
|
|
|
172 fp = CURSTR();
|
|
|
173
|
|
|
174 // fp is the symbol name
|
|
|
175 val = (CURBYTE()) << 8;
|
|
|
176 NEXTBYTE();
|
|
|
177 val |= (CURBYTE());
|
|
|
178 NEXTBYTE();
|
|
|
179 // val is now the symbol value
|
|
|
180
|
|
|
181 // create symbol table entry
|
|
|
182 se = lw_malloc(sizeof(symtab_t));
|
|
|
183 se -> next = s -> localsyms;
|
|
|
184 s -> localsyms = se;
|
|
|
185 se -> sym = fp;
|
|
|
186 se -> offset = val;
|
|
|
187 }
|
|
124
|
188 // skip terminating NUL
|
|
|
189 NEXTBYTE();
|
|
|
190
|
|
116
|
191 // now parse the exported symbol table
|
|
|
192 while (CURBYTE())
|
|
|
193 {
|
|
|
194 fp = CURSTR();
|
|
|
195
|
|
|
196 // fp is the symbol name
|
|
|
197 val = (CURBYTE()) << 8;
|
|
|
198 NEXTBYTE();
|
|
|
199 val |= (CURBYTE());
|
|
|
200 NEXTBYTE();
|
|
|
201 // val is now the symbol value
|
|
|
202
|
|
|
203 // create symbol table entry
|
|
|
204 se = lw_malloc(sizeof(symtab_t));
|
|
|
205 se -> next = s -> exportedsyms;
|
|
|
206 s -> localsyms = se;
|
|
|
207 se -> sym = fp;
|
|
|
208 se -> offset = val;
|
|
|
209 }
|
|
124
|
210 // skip terminating NUL
|
|
|
211 NEXTBYTE();
|
|
116
|
212
|
|
|
213 // now parse the incomplete references and make a list of
|
|
|
214 // external references that need resolution
|
|
|
215 while (CURBYTE())
|
|
|
216 {
|
|
|
217 reloc_t *rp;
|
|
|
218 lw_expr_term_t *term;
|
|
|
219
|
|
|
220 // we have a reference
|
|
|
221 rp = lw_malloc(sizeof(reloc_t));
|
|
|
222 rp -> next = s -> incompletes;
|
|
|
223 s -> incompletes = rp;
|
|
|
224 rp -> offset = 0;
|
|
|
225 rp -> expr = lw_expr_stack_create();
|
|
|
226
|
|
|
227 // parse the expression
|
|
|
228 while (CURBYTE())
|
|
|
229 {
|
|
|
230 int tt = CURBYTE();
|
|
|
231 NEXTBYTE();
|
|
|
232 switch (tt)
|
|
|
233 {
|
|
|
234 case 0x01:
|
|
|
235 // 16 bit integer
|
|
|
236 tt = CURBYTE() << 8;
|
|
|
237 NEXTBYTE();
|
|
|
238 tt |= CURBYTE();
|
|
|
239 NEXTBYTE();
|
|
|
240 // normalize for negatives...
|
|
|
241 if (tt > 0x7fff)
|
|
|
242 tt -= 0x10000;
|
|
|
243 term = lw_expr_term_create_int(tt);
|
|
|
244 break;
|
|
|
245
|
|
|
246 case 0x02:
|
|
|
247 // external symbol reference
|
|
|
248 term = lw_expr_term_create_sym(CURSTR(), 0);
|
|
|
249 break;
|
|
|
250
|
|
|
251 case 0x03:
|
|
|
252 // internal symbol reference
|
|
|
253 term = lw_expr_term_create_sym(CURSTR(), 1);
|
|
|
254 break;
|
|
|
255
|
|
|
256 case 0x04:
|
|
|
257 // operator
|
|
|
258 term = lw_expr_term_create_oper(CURBYTE());
|
|
|
259 NEXTBYTE();
|
|
|
260 break;
|
|
|
261
|
|
|
262 case 0x05:
|
|
|
263 // section base reference (NULL internal reference is
|
|
|
264 // the section base address
|
|
120
|
265 term = lw_expr_term_create_sym(NULL, 1);
|
|
116
|
266 break;
|
|
|
267
|
|
|
268 default:
|
|
|
269 fprintf(stderr, "%s (%s): bad relocation expression\n", fn -> filename, s -> name);
|
|
|
270 exit(1);
|
|
|
271 }
|
|
|
272 lw_expr_stack_push(rp -> expr, term);
|
|
|
273 lw_expr_term_free(term);
|
|
|
274 }
|
|
127
|
275 // skip the NUL
|
|
|
276 NEXTBYTE();
|
|
|
277
|
|
|
278 // fetch the offset
|
|
|
279 rp -> offset = CURBYTE() << 8;
|
|
|
280 NEXTBYTE();
|
|
|
281 rp -> offset |= CURBYTE() & 0xff;
|
|
|
282 NEXTBYTE();
|
|
116
|
283 }
|
|
124
|
284 // skip the NUL terminating the relocations
|
|
116
|
285 NEXTBYTE();
|
|
|
286
|
|
|
287 // now set code location and size and verify that the file
|
|
|
288 // contains data going to the end of the code (if !SECTION_BSS)
|
|
|
289 s -> codesize = CURBYTE() << 8;
|
|
|
290 NEXTBYTE();
|
|
|
291 s -> codesize |= CURBYTE();
|
|
|
292 NEXTBYTE();
|
|
|
293
|
|
|
294 s -> code = &(CURBYTE());
|
|
|
295
|
|
|
296 // skip the code if we're not in a BSS section
|
|
|
297 if (!(s -> flags & SECTION_BSS))
|
|
|
298 {
|
|
|
299 int i;
|
|
|
300 for (i = 0; i < s -> codesize; i++)
|
|
|
301 NEXTBYTE();
|
|
|
302 }
|
|
|
303 }
|
|
|
304 }
|
