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>
|
171
|
33 #include <string.h>
|
116
|
34
|
|
35 #include "lwlink.h"
|
|
36 #include "util.h"
|
|
37
|
|
38 void read_lwobj16v0(fileinfo_t *fn);
|
171
|
39 void read_lwar1v(fileinfo_t *fn);
|
116
|
40
|
|
41 /*
|
|
42 The logic of reading the entire file into memory is simple. All the symbol
|
|
43 names in the file are NUL terminated strings and can be used directly without
|
|
44 making additional copies.
|
|
45 */
|
171
|
46 void read_file(fileinfo_t *fn)
|
|
47 {
|
|
48 if (!memcmp(fn -> filedata, "LWOBJ16", 8))
|
|
49 {
|
|
50 // read v0 LWOBJ16 file
|
|
51 read_lwobj16v0(fn);
|
|
52 }
|
|
53 else if (!memcmp(fn -> filedata, "LWAR1V", 6))
|
|
54 {
|
|
55 // archive file
|
|
56 read_lwar1v(fn);
|
|
57 }
|
|
58 else
|
|
59 {
|
|
60 fprintf(stderr, "%s: unknown file format\n", fn -> filename);
|
|
61 exit(1);
|
|
62 }
|
|
63 }
|
|
64
|
116
|
65 void read_files(void)
|
|
66 {
|
|
67 int i;
|
|
68 long size;
|
|
69 FILE *f;
|
|
70 long bread;
|
|
71 for (i = 0; i < ninputfiles; i++)
|
|
72 {
|
|
73 f = fopen(inputfiles[i] -> filename, "rb");
|
|
74 if (!f)
|
|
75 {
|
|
76 fprintf(stderr, "Can't open file %s:", inputfiles[i] -> filename);
|
|
77 perror("");
|
|
78 exit(1);
|
|
79 }
|
|
80 fseek(f, 0, SEEK_END);
|
|
81 size = ftell(f);
|
|
82 rewind(f);
|
|
83
|
|
84 inputfiles[i] -> filedata = lw_malloc(size);
|
|
85 inputfiles[i] -> filesize = size;
|
|
86
|
|
87 bread = fread(inputfiles[i] -> filedata, 1, size, f);
|
|
88 if (bread < size)
|
|
89 {
|
|
90 fprintf(stderr, "Short read on file %s (%ld/%ld):", inputfiles[i] -> filename, bread, size);
|
|
91 perror("");
|
|
92 exit(1);
|
|
93 }
|
|
94
|
|
95 fclose(f);
|
|
96
|
171
|
97 read_file(inputfiles[i]);
|
116
|
98 }
|
|
99 }
|
|
100
|
|
101 // this macro is used to bail out if we run off the end of the file data
|
|
102 // while parsing - it keeps the code below cleaner
|
124
|
103 #define NEXTBYTE() do { cc++; if (cc > fn -> filesize) { fprintf(stderr, "%s: invalid file format\n", fn -> filename); exit(1); } } while (0)
|
116
|
104 // this macro is used to refer to the current byte in the stream
|
124
|
105 #define CURBYTE() (fn -> filedata[cc < fn -> filesize ? cc : fn -> filesize - 1])
|
116
|
106 // this one will leave the input pointer past the trailing NUL
|
|
107 #define CURSTR() read_lwobj16v0_str(&cc, fn)
|
|
108 unsigned char *read_lwobj16v0_str(long *cc1, fileinfo_t *fn)
|
|
109 {
|
|
110 int cc = *cc1;
|
|
111 unsigned char *fp;
|
|
112 fp = &CURBYTE();
|
|
113 while (CURBYTE())
|
|
114 NEXTBYTE();
|
|
115 NEXTBYTE();
|
|
116 *cc1 = cc;
|
|
117 return fp;
|
|
118 }
|
|
119 // the function below can be switched to dealing with data coming from a
|
|
120 // source other than an in-memory byte pool by adjusting the input data
|
|
121 // in "fn" and the above two macros
|
171
|
122
|
116
|
123 void read_lwobj16v0(fileinfo_t *fn)
|
|
124 {
|
|
125 unsigned char *fp;
|
|
126 long cc;
|
|
127 section_t *s;
|
|
128 int val;
|
|
129 symtab_t *se;
|
|
130
|
|
131 // start reading *after* the magic number
|
126
|
132 cc = 8;
|
116
|
133
|
|
134 // init data
|
|
135 fn -> sections = NULL;
|
|
136 fn -> nsections = 0;
|
|
137
|
|
138 while (1)
|
|
139 {
|
126
|
140 // NEXTBYTE();
|
116
|
141 // bail out if no more sections
|
|
142 if (!(CURBYTE()))
|
|
143 break;
|
|
144
|
|
145 fp = CURSTR();
|
|
146
|
|
147 // we now have a section name in fp
|
|
148 // create new section entry
|
|
149 fn -> sections = lw_realloc(fn -> sections, sizeof(section_t) * (fn -> nsections + 1));
|
|
150 s = &(fn -> sections[fn -> nsections]);
|
|
151 fn -> nsections += 1;
|
|
152
|
|
153 s -> localsyms = NULL;
|
|
154 s -> flags = 0;
|
|
155 s -> codesize = 0;
|
|
156 s -> name = fp;
|
|
157 s -> loadaddress = 0;
|
|
158 s -> localsyms = NULL;
|
|
159 s -> exportedsyms = NULL;
|
|
160 s -> incompletes = NULL;
|
119
|
161 s -> processed = 0;
|
|
162 s -> file = fn;
|
|
163
|
116
|
164 // read flags
|
|
165 while (CURBYTE())
|
|
166 {
|
|
167 switch (CURBYTE())
|
|
168 {
|
|
169 case 0x01:
|
|
170 s -> flags |= SECTION_BSS;
|
|
171 break;
|
|
172
|
|
173 default:
|
|
174 fprintf(stderr, "%s (%s): unrecognized section flag %02X\n", fn -> filename, s -> name, (int)(CURBYTE()));
|
|
175 exit(1);
|
|
176 }
|
|
177 NEXTBYTE();
|
|
178 }
|
|
179 // skip NUL terminating flags
|
|
180 NEXTBYTE();
|
|
181
|
|
182 // now parse the local symbol table
|
|
183 while (CURBYTE())
|
|
184 {
|
|
185 fp = CURSTR();
|
|
186
|
|
187 // fp is the symbol name
|
|
188 val = (CURBYTE()) << 8;
|
|
189 NEXTBYTE();
|
|
190 val |= (CURBYTE());
|
|
191 NEXTBYTE();
|
|
192 // val is now the symbol value
|
|
193
|
|
194 // create symbol table entry
|
|
195 se = lw_malloc(sizeof(symtab_t));
|
|
196 se -> next = s -> localsyms;
|
|
197 s -> localsyms = se;
|
|
198 se -> sym = fp;
|
|
199 se -> offset = val;
|
|
200 }
|
124
|
201 // skip terminating NUL
|
|
202 NEXTBYTE();
|
|
203
|
116
|
204 // now parse the exported symbol table
|
|
205 while (CURBYTE())
|
|
206 {
|
|
207 fp = CURSTR();
|
|
208
|
|
209 // fp is the symbol name
|
|
210 val = (CURBYTE()) << 8;
|
|
211 NEXTBYTE();
|
|
212 val |= (CURBYTE());
|
|
213 NEXTBYTE();
|
|
214 // val is now the symbol value
|
|
215
|
|
216 // create symbol table entry
|
|
217 se = lw_malloc(sizeof(symtab_t));
|
|
218 se -> next = s -> exportedsyms;
|
128
|
219 s -> exportedsyms = se;
|
116
|
220 se -> sym = fp;
|
|
221 se -> offset = val;
|
|
222 }
|
124
|
223 // skip terminating NUL
|
|
224 NEXTBYTE();
|
116
|
225
|
|
226 // now parse the incomplete references and make a list of
|
|
227 // external references that need resolution
|
|
228 while (CURBYTE())
|
|
229 {
|
|
230 reloc_t *rp;
|
|
231 lw_expr_term_t *term;
|
|
232
|
|
233 // we have a reference
|
|
234 rp = lw_malloc(sizeof(reloc_t));
|
|
235 rp -> next = s -> incompletes;
|
|
236 s -> incompletes = rp;
|
|
237 rp -> offset = 0;
|
|
238 rp -> expr = lw_expr_stack_create();
|
|
239
|
|
240 // parse the expression
|
|
241 while (CURBYTE())
|
|
242 {
|
|
243 int tt = CURBYTE();
|
|
244 NEXTBYTE();
|
|
245 switch (tt)
|
|
246 {
|
|
247 case 0x01:
|
|
248 // 16 bit integer
|
|
249 tt = CURBYTE() << 8;
|
|
250 NEXTBYTE();
|
|
251 tt |= CURBYTE();
|
|
252 NEXTBYTE();
|
|
253 // normalize for negatives...
|
|
254 if (tt > 0x7fff)
|
|
255 tt -= 0x10000;
|
|
256 term = lw_expr_term_create_int(tt);
|
|
257 break;
|
|
258
|
|
259 case 0x02:
|
|
260 // external symbol reference
|
|
261 term = lw_expr_term_create_sym(CURSTR(), 0);
|
|
262 break;
|
|
263
|
|
264 case 0x03:
|
|
265 // internal symbol reference
|
|
266 term = lw_expr_term_create_sym(CURSTR(), 1);
|
|
267 break;
|
|
268
|
|
269 case 0x04:
|
|
270 // operator
|
|
271 term = lw_expr_term_create_oper(CURBYTE());
|
|
272 NEXTBYTE();
|
|
273 break;
|
|
274
|
|
275 case 0x05:
|
|
276 // section base reference (NULL internal reference is
|
|
277 // the section base address
|
120
|
278 term = lw_expr_term_create_sym(NULL, 1);
|
116
|
279 break;
|
|
280
|
|
281 default:
|
|
282 fprintf(stderr, "%s (%s): bad relocation expression\n", fn -> filename, s -> name);
|
|
283 exit(1);
|
|
284 }
|
|
285 lw_expr_stack_push(rp -> expr, term);
|
|
286 lw_expr_term_free(term);
|
|
287 }
|
127
|
288 // skip the NUL
|
|
289 NEXTBYTE();
|
|
290
|
|
291 // fetch the offset
|
|
292 rp -> offset = CURBYTE() << 8;
|
|
293 NEXTBYTE();
|
|
294 rp -> offset |= CURBYTE() & 0xff;
|
|
295 NEXTBYTE();
|
116
|
296 }
|
124
|
297 // skip the NUL terminating the relocations
|
116
|
298 NEXTBYTE();
|
|
299
|
|
300 // now set code location and size and verify that the file
|
|
301 // contains data going to the end of the code (if !SECTION_BSS)
|
|
302 s -> codesize = CURBYTE() << 8;
|
|
303 NEXTBYTE();
|
|
304 s -> codesize |= CURBYTE();
|
|
305 NEXTBYTE();
|
|
306
|
|
307 s -> code = &(CURBYTE());
|
|
308
|
|
309 // skip the code if we're not in a BSS section
|
|
310 if (!(s -> flags & SECTION_BSS))
|
|
311 {
|
|
312 int i;
|
|
313 for (i = 0; i < s -> codesize; i++)
|
|
314 NEXTBYTE();
|
|
315 }
|
|
316 }
|
|
317 }
|
171
|
318
|
|
319 /*
|
|
320 Read an archive file - this will create a "sub" record and farm out the
|
|
321 parsing of the sub files to the regular file parsers
|
|
322
|
|
323 The archive file format consists of the 6 byte magic number followed by a
|
|
324 series of records as follows:
|
|
325
|
|
326 - NUL terminated file name
|
|
327 - 32 bit file length in big endian order
|
|
328 - the file data
|
|
329
|
|
330 An empty file name indicates the end of the file.
|
|
331
|
|
332 */
|
|
333 void read_lwar1v(fileinfo_t *fn)
|
|
334 {
|
|
335 unsigned long cc = 6;
|
|
336 unsigned long flen;
|
|
337 unsigned long l;
|
|
338 for (;;)
|
|
339 {
|
|
340 if (cc >= fn -> filesize || !(fn -> filedata[cc]))
|
|
341 return;
|
|
342
|
|
343 for (l = cc; cc < fn -> filesize && fn -> filedata[cc]; l++)
|
|
344 /* do nothing */ ;
|
|
345
|
|
346 if (cc >= fn -> filesize)
|
|
347 {
|
|
348 fprintf(stderr, "Malformed archive file %s.\n", fn -> filename);
|
|
349 exit(1);
|
|
350 }
|
|
351
|
|
352 if (cc + 4 > fn -> filesize)
|
|
353 return;
|
|
354
|
|
355 flen = (fn -> filedata[cc++] << 24) | (fn -> filedata[cc++] << 16)
|
|
356 | (fn -> filedata[cc++] << 8) | (fn -> filedata[cc]);
|
|
357
|
|
358 if (flen == 0)
|
|
359 return;
|
|
360
|
|
361 if (cc + flen > fn -> filesize)
|
|
362 {
|
|
363 fprintf(stderr, "Malformed archive file %s.\n", fn -> filename);
|
|
364 exit(1);
|
|
365 }
|
|
366
|
|
367 // add the "sub" input file
|
|
368 fn -> subs = lw_realloc(fn -> subs, sizeof(fileinfo_t *) * (fn -> nsubs + 1));
|
|
369 fn -> subs[fn -> nsubs] = lw_malloc(sizeof(fileinfo_t));
|
|
370 memset(fn -> subs[fn -> nsubs], 0, sizeof(fileinfo_t));
|
|
371 fn -> subs[fn -> nsubs] -> filedata = fn -> filedata + cc;
|
|
372 fn -> subs[fn -> nsubs] -> filesize = flen;
|
|
373 fn -> subs[fn -> nsubs] -> filename = lw_strdup(fn -> filedata + l);
|
|
374 fn -> subs[fn -> nsubs] -> parent = fn;
|
|
375 read_file(fn -> subs[fn -> nsubs]);
|
|
376 fn -> nsubs++;
|
|
377 }
|
|
378 }
|