1 /*
2 * CDDL HEADER START
3 *
4 * The contents of this file are subject to the terms of the
5 * Common Development and Distribution License (the "License").
6 * You may not use this file except in compliance with the License.
7 *
8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 * or http://www.opensolaris.org/os/licensing.
10 * See the License for the specific language governing permissions
11 * and limitations under the License.
12 *
13 * When distributing Covered Code, include this CDDL HEADER in each
14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 * If applicable, add the following below this CDDL HEADER, with the
16 * fields enclosed by brackets "[]" replaced with your own identifying
17 * information: Portions Copyright [yyyy] [name of copyright owner]
18 *
19 * CDDL HEADER END
20 */
21 /*
22 * Copyright 2006 Sun Microsystems, Inc. All rights reserved.
23 * Use is subject to license terms.
24 */
25
26 /*
27 * Routines for preparing tdata trees for conversion into CTF data, and
28 * for placing the resulting data into an output file.
29 */
30
31 #include <stdio.h>
32 #include <stdlib.h>
33 #include <strings.h>
34 #include <sys/types.h>
35 #include <sys/stat.h>
36 #include <fcntl.h>
37 #include <libelf.h>
38 #include <gelf.h>
39 #include <unistd.h>
40
41 #include "ctftools.h"
42 #include "list.h"
43 #include "memory.h"
44 #include "traverse.h"
45 #include "symbol.h"
46
47 typedef struct iidesc_match {
48 int iim_fuzzy;
49 iidesc_t *iim_ret;
50 char *iim_name;
51 char *iim_file;
52 uchar_t iim_bind;
53 } iidesc_match_t;
54
55 static int
56 burst_iitypes(void *data, void *arg)
57 {
58 iidesc_t *ii = data;
59 iiburst_t *iiburst = arg;
60
61 switch (ii->ii_type) {
62 case II_GFUN:
63 case II_SFUN:
64 case II_GVAR:
65 case II_SVAR:
66 if (!(ii->ii_flags & IIDESC_F_USED))
67 return (0);
68 break;
69 default:
70 break;
71 }
72
73 ii->ii_dtype->t_flags |= TDESC_F_ISROOT;
74 (void) iitraverse_td(ii, iiburst->iib_tdtd);
75 return (1);
76 }
77
78 /*ARGSUSED1*/
79 static int
80 save_type_by_id(tdesc_t *tdp, tdesc_t **tdpp, void *private)
81 {
82 iiburst_t *iiburst = private;
83
84 /*
85 * Doing this on every node is horribly inefficient, but given that
86 * we may be suppressing some types, we can't trust nextid in the
87 * tdata_t.
88 */
89 if (tdp->t_id > iiburst->iib_maxtypeid)
90 iiburst->iib_maxtypeid = tdp->t_id;
91
92 slist_add(&iiburst->iib_types, tdp, tdesc_idcmp);
93
94 return (1);
95 }
96
97 static tdtrav_cb_f burst_types_cbs[] = {
98 NULL,
99 save_type_by_id, /* intrinsic */
100 save_type_by_id, /* pointer */
101 save_type_by_id, /* array */
102 save_type_by_id, /* function */
103 save_type_by_id, /* struct */
104 save_type_by_id, /* union */
105 save_type_by_id, /* enum */
106 save_type_by_id, /* forward */
107 save_type_by_id, /* typedef */
108 tdtrav_assert, /* typedef_unres */
109 save_type_by_id, /* volatile */
110 save_type_by_id, /* const */
111 save_type_by_id /* restrict */
112 };
113
114
115 static iiburst_t *
116 iiburst_new(tdata_t *td, int max)
117 {
118 iiburst_t *iiburst = xcalloc(sizeof (iiburst_t));
119 iiburst->iib_td = td;
120 iiburst->iib_funcs = xcalloc(sizeof (iidesc_t *) * max);
121 iiburst->iib_nfuncs = 0;
122 iiburst->iib_objts = xcalloc(sizeof (iidesc_t *) * max);
123 iiburst->iib_nobjts = 0;
124 return (iiburst);
125 }
126
127 static void
128 iiburst_types(iiburst_t *iiburst)
129 {
130 tdtrav_data_t tdtd;
131
132 tdtrav_init(&tdtd, &iiburst->iib_td->td_curvgen, NULL, burst_types_cbs,
133 NULL, (void *)iiburst);
134
135 iiburst->iib_tdtd = &tdtd;
136
137 (void) hash_iter(iiburst->iib_td->td_iihash, burst_iitypes, iiburst);
138 }
139
140 static void
141 iiburst_free(iiburst_t *iiburst)
142 {
143 free(iiburst->iib_funcs);
144 free(iiburst->iib_objts);
145 list_free(iiburst->iib_types, NULL, NULL);
146 free(iiburst);
147 }
148
149 /*
150 * See if this iidesc matches the ELF symbol data we pass in.
151 *
152 * A fuzzy match is where we have a local symbol matching the name of a
153 * global type description. This is common when a mapfile is used for a
154 * DSO, but we don't accept it by default.
155 *
156 * A weak fuzzy match is when a weak symbol was resolved and matched to
157 * a global type description.
158 */
159 static int
160 matching_iidesc(iidesc_t *iidesc, iidesc_match_t *match)
161 {
162 if (streq(iidesc->ii_name, match->iim_name) == 0)
163 return (0);
164
165 switch (iidesc->ii_type) {
166 case II_GFUN:
167 case II_GVAR:
168 if (match->iim_bind == STB_GLOBAL) {
169 match->iim_ret = iidesc;
170 return (-1);
171 } else if (match->iim_fuzzy && match->iim_ret == NULL) {
172 match->iim_ret = iidesc;
173 /* continue to look for strong match */
174 return (0);
175 }
176 break;
177 case II_SFUN:
178 case II_SVAR:
179 if (match->iim_bind == STB_LOCAL &&
180 match->iim_file != NULL &&
181 streq(iidesc->ii_owner, match->iim_file)) {
182 match->iim_ret = iidesc;
183 return (-1);
184 }
185 break;
186 }
187 return (0);
188 }
189
190 static iidesc_t *
191 find_iidesc(tdata_t *td, iidesc_match_t *match)
192 {
193 match->iim_ret = NULL;
194 iter_iidescs_by_name(td, match->iim_name,
195 (int (*)())matching_iidesc, match);
196 return (match->iim_ret);
197 }
198
199 /*
200 * If we have a weak symbol, attempt to find the strong symbol it will
201 * resolve to. Note: the code where this actually happens is in
202 * sym_process() in cmd/sgs/libld/common/syms.c
203 *
204 * Finding the matching symbol is unfortunately not trivial. For a
205 * symbol to be a candidate, it must:
206 *
207 * - have the same type (function, object)
208 * - have the same value (address)
209 * - have the same size
210 * - not be another weak symbol
211 * - belong to the same section (checked via section index)
212 *
213 * If such a candidate is global, then we assume we've found it. The
214 * linker generates the symbol table such that the curfile might be
215 * incorrect; this is OK for global symbols, since find_iidesc() doesn't
216 * need to check for the source file for the symbol.
217 *
218 * We might have found a strong local symbol, where the curfile is
219 * accurate and matches that of the weak symbol. We assume this is a
220 * reasonable match.
221 *
222 * If we've got a local symbol with a non-matching curfile, there are
223 * two possibilities. Either this is a completely different symbol, or
224 * it's a once-global symbol that was scoped to local via a mapfile. In
225 * the latter case, curfile is likely inaccurate since the linker does
226 * not preserve the needed curfile in the order of the symbol table (see
227 * the comments about locally scoped symbols in libld's update_osym()).
228 * As we can't tell this case from the former one, we use this symbol
229 * iff no other matching symbol is found.
230 *
231 * What we really need here is a SUNW section containing weak<->strong
232 * mappings that we can consume.
233 */
234 static int
235 check_for_weak(GElf_Sym *weak, char const *weakfile,
236 Elf_Data *data, int nent, Elf_Data *strdata,
237 GElf_Sym *retsym, char **curfilep)
238 {
239 char *curfile = NULL;
240 char *tmpfile;
241 GElf_Sym tmpsym;
242 int candidate = 0;
243 int i;
244
245 if (GELF_ST_BIND(weak->st_info) != STB_WEAK)
246 return (0);
247
248 for (i = 0; i < nent; i++) {
249 GElf_Sym sym;
250 uchar_t type;
251
252 if (gelf_getsym(data, i, &sym) == NULL)
253 continue;
254
255 type = GELF_ST_TYPE(sym.st_info);
256
257 if (type == STT_FILE)
258 curfile = (char *)strdata->d_buf + sym.st_name;
259
260 if (GELF_ST_TYPE(weak->st_info) != type ||
261 weak->st_value != sym.st_value)
262 continue;
263
264 if (weak->st_size != sym.st_size)
265 continue;
266
267 if (GELF_ST_BIND(sym.st_info) == STB_WEAK)
268 continue;
269
270 if (sym.st_shndx != weak->st_shndx)
271 continue;
272
273 if (GELF_ST_BIND(sym.st_info) == STB_LOCAL &&
274 (curfile == NULL || weakfile == NULL ||
275 strcmp(curfile, weakfile) != 0)) {
276 candidate = 1;
277 tmpfile = curfile;
278 tmpsym = sym;
279 continue;
280 }
281
282 *curfilep = curfile;
283 *retsym = sym;
284 return (1);
285 }
286
287 if (candidate) {
288 *curfilep = tmpfile;
289 *retsym = tmpsym;
290 return (1);
291 }
292
293 return (0);
294 }
295
296 /*
297 * When we've found the underlying symbol's type description
298 * for a weak symbol, we need to copy it and rename it to match
299 * the weak symbol. We also need to add it to the td so it's
300 * handled along with the others later.
301 */
302 static iidesc_t *
303 copy_from_strong(tdata_t *td, GElf_Sym *sym, iidesc_t *strongdesc,
304 const char *weakname, const char *weakfile)
305 {
306 iidesc_t *new = iidesc_dup_rename(strongdesc, weakname, weakfile);
307 uchar_t type = GELF_ST_TYPE(sym->st_info);
308
309 switch (type) {
310 case STT_OBJECT:
311 new->ii_type = II_GVAR;
312 break;
313 case STT_FUNC:
314 new->ii_type = II_GFUN;
315 break;
316 }
317
318 hash_add(td->td_iihash, new);
319
320 return (new);
321 }
322
323 /*
324 * Process the symbol table of the output file, associating each symbol
325 * with a type description if possible, and sorting them into functions
326 * and data, maintaining symbol table order.
327 */
328 static iiburst_t *
329 sort_iidescs(Elf *elf, const char *file, tdata_t *td, int fuzzymatch,
330 int dynsym)
331 {
332 iiburst_t *iiburst;
333 Elf_Scn *scn;
334 GElf_Shdr shdr;
335 Elf_Data *data, *strdata;
336 int i, stidx;
337 int nent;
338 iidesc_match_t match;
339
340 match.iim_fuzzy = fuzzymatch;
341 match.iim_file = NULL;
342
343 if ((stidx = findelfsecidx(elf, file,
344 dynsym ? ".dynsym" : ".symtab")) < 0)
345 terminate("%s: Can't open symbol table\n", file);
346 scn = elf_getscn(elf, stidx);
347 data = elf_getdata(scn, NULL);
348 gelf_getshdr(scn, &shdr);
349 nent = shdr.sh_size / shdr.sh_entsize;
350
351 scn = elf_getscn(elf, shdr.sh_link);
352 strdata = elf_getdata(scn, NULL);
353
354 iiburst = iiburst_new(td, nent);
355
356 for (i = 0; i < nent; i++) {
357 GElf_Sym sym;
358 iidesc_t **tolist;
359 GElf_Sym ssym;
360 iidesc_match_t smatch;
361 int *curr;
362 iidesc_t *iidesc;
363
364 if (gelf_getsym(data, i, &sym) == NULL)
365 elfterminate(file, "Couldn't read symbol %d", i);
366
367 match.iim_name = (char *)strdata->d_buf + sym.st_name;
368 match.iim_bind = GELF_ST_BIND(sym.st_info);
369
370 switch (GELF_ST_TYPE(sym.st_info)) {
371 case STT_FILE:
372 match.iim_file = match.iim_name;
373 continue;
374 case STT_OBJECT:
375 tolist = iiburst->iib_objts;
376 curr = &iiburst->iib_nobjts;
377 break;
378 case STT_FUNC:
379 tolist = iiburst->iib_funcs;
380 curr = &iiburst->iib_nfuncs;
381 break;
382 default:
383 continue;
384 }
385
386 if (ignore_symbol(&sym, match.iim_name))
387 continue;
388
389 iidesc = find_iidesc(td, &match);
390
391 if (iidesc != NULL) {
392 tolist[*curr] = iidesc;
393 iidesc->ii_flags |= IIDESC_F_USED;
394 (*curr)++;
395 continue;
396 }
397
398 if (!check_for_weak(&sym, match.iim_file, data, nent, strdata,
399 &ssym, &smatch.iim_file)) {
400 (*curr)++;
401 continue;
402 }
403
404 smatch.iim_fuzzy = fuzzymatch;
405 smatch.iim_name = (char *)strdata->d_buf + ssym.st_name;
406 smatch.iim_bind = GELF_ST_BIND(ssym.st_info);
407
408 debug(3, "Weak symbol %s resolved to %s\n", match.iim_name,
409 smatch.iim_name);
410
411 iidesc = find_iidesc(td, &smatch);
412
413 if (iidesc != NULL) {
414 tolist[*curr] = copy_from_strong(td, &sym,
415 iidesc, match.iim_name, match.iim_file);
416 tolist[*curr]->ii_flags |= IIDESC_F_USED;
417 }
418
419 (*curr)++;
420 }
421
422 /*
423 * Stabs are generated for every function declared in a given C source
424 * file. When converting an object file, we may encounter a stab that
425 * has no symbol table entry because the optimizer has decided to omit
426 * that item (for example, an unreferenced static function). We may
427 * see iidescs that do not have an associated symtab entry, and so
428 * we do not write records for those functions into the CTF data.
429 * All others get marked as a root by this function.
430 */
431 iiburst_types(iiburst);
432
433 /*
434 * By not adding some of the functions and/or objects, we may have
435 * caused some types that were referenced solely by those
436 * functions/objects to be suppressed. This could cause a label,
437 * generated prior to the evisceration, to be incorrect. Find the
438 * highest type index, and change the label indicies to be no higher
439 * than this value.
440 */
441 tdata_label_newmax(td, iiburst->iib_maxtypeid);
442
443 return (iiburst);
444 }
445
446 static void
447 write_file(Elf *src, const char *srcname, Elf *dst, const char *dstname,
448 caddr_t ctfdata, size_t ctfsize, int flags)
449 {
450 GElf_Ehdr sehdr, dehdr;
451 Elf_Scn *sscn, *dscn;
452 Elf_Data *sdata, *ddata;
453 GElf_Shdr shdr;
454 GElf_Word symtab_type;
455 int symtab_idx = -1;
456 off_t new_offset = 0;
457 off_t ctfnameoff = 0;
458 int dynsym = (flags & CTF_USE_DYNSYM);
459 int *secxlate;
460 int srcidx, dstidx;
461 int curnmoff = 0;
462 int changing = 0;
463 int pad;
464 int i;
465
466 if (gelf_newehdr(dst, gelf_getclass(src)) == 0)
467 elfterminate(dstname, "Cannot copy ehdr to temp file");
468 gelf_getehdr(src, &sehdr);
469 memcpy(&dehdr, &sehdr, sizeof (GElf_Ehdr));
470 gelf_update_ehdr(dst, &dehdr);
471
472 symtab_type = dynsym ? SHT_DYNSYM : SHT_SYMTAB;
473
474 /*
475 * Neither the existing stab sections nor the SUNW_ctf sections (new or
476 * existing) are SHF_ALLOC'd, so they won't be in areas referenced by
477 * program headers. As such, we can just blindly copy the program
478 * headers from the existing file to the new file.
479 */
480 if (sehdr.e_phnum != 0) {
481 (void) elf_flagelf(dst, ELF_C_SET, ELF_F_LAYOUT);
482 if (gelf_newphdr(dst, sehdr.e_phnum) == 0)
483 elfterminate(dstname, "Cannot make phdrs in temp file");
484
485 for (i = 0; i < sehdr.e_phnum; i++) {
486 GElf_Phdr phdr;
487
488 gelf_getphdr(src, i, &phdr);
489 gelf_update_phdr(dst, i, &phdr);
490 }
491 }
492
493 secxlate = xmalloc(sizeof (int) * sehdr.e_shnum);
494 for (srcidx = dstidx = 0; srcidx < sehdr.e_shnum; srcidx++) {
495 Elf_Scn *scn = elf_getscn(src, srcidx);
496 GElf_Shdr shdr;
497 char *sname;
498
499 gelf_getshdr(scn, &shdr);
500 sname = elf_strptr(src, sehdr.e_shstrndx, shdr.sh_name);
501 if (sname == NULL) {
502 elfterminate(srcname, "Can't find string at %u",
503 shdr.sh_name);
504 }
505
506 if (strcmp(sname, CTF_ELF_SCN_NAME) == 0) {
507 secxlate[srcidx] = -1;
508 } else if (dynsym && shdr.sh_type == SHT_SYMTAB) {
509 /*
510 * If we're building CTF against the dynsym,
511 * we'll rip out the symtab so debuggers aren't
512 * confused.
513 */
514 secxlate[srcidx] = -1;
515 } else {
516 secxlate[srcidx] = dstidx++;
517 curnmoff += strlen(sname) + 1;
518 }
519
520 new_offset = (off_t)dehdr.e_phoff;
521 }
522
523 for (srcidx = 1; srcidx < sehdr.e_shnum; srcidx++) {
524 char *sname;
525
526 sscn = elf_getscn(src, srcidx);
527 gelf_getshdr(sscn, &shdr);
528
529 if (secxlate[srcidx] == -1) {
530 changing = 1;
531 continue;
532 }
533
534 dscn = elf_newscn(dst);
535
536 /*
537 * If this file has program headers, we need to explicitly lay
538 * out sections. If none of the sections prior to this one have
539 * been removed, then we can just use the existing location. If
540 * one or more sections have been changed, then we need to
541 * adjust this one to avoid holes.
542 */
543 if (changing && sehdr.e_phnum != 0) {
544 pad = new_offset % shdr.sh_addralign;
545
546 if (pad)
547 new_offset += shdr.sh_addralign - pad;
548 shdr.sh_offset = new_offset;
549 }
550
551 shdr.sh_link = secxlate[shdr.sh_link];
552
553 if (shdr.sh_type == SHT_REL || shdr.sh_type == SHT_RELA)
554 shdr.sh_info = secxlate[shdr.sh_info];
555
556 sname = elf_strptr(src, sehdr.e_shstrndx, shdr.sh_name);
557 if (sname == NULL) {
558 elfterminate(srcname, "Can't find string at %u",
559 shdr.sh_name);
560 }
561 if ((sdata = elf_getdata(sscn, NULL)) == NULL)
562 elfterminate(srcname, "Cannot get sect %s data", sname);
563 if ((ddata = elf_newdata(dscn)) == NULL)
564 elfterminate(dstname, "Can't make sect %s data", sname);
565 bcopy(sdata, ddata, sizeof (Elf_Data));
566
567 if (srcidx == sehdr.e_shstrndx) {
568 char seclen = strlen(CTF_ELF_SCN_NAME);
569
570 ddata->d_buf = xmalloc(ddata->d_size + shdr.sh_size +
571 seclen + 1);
572 bcopy(sdata->d_buf, ddata->d_buf, shdr.sh_size);
573 strcpy((caddr_t)ddata->d_buf + shdr.sh_size,
574 CTF_ELF_SCN_NAME);
575 ctfnameoff = (off_t)shdr.sh_size;
576 shdr.sh_size += seclen + 1;
577 ddata->d_size += seclen + 1;
578
579 if (sehdr.e_phnum != 0)
580 changing = 1;
581 }
582
583 if (shdr.sh_type == symtab_type && shdr.sh_entsize != 0) {
584 int nsym = shdr.sh_size / shdr.sh_entsize;
585
586 symtab_idx = secxlate[srcidx];
587
588 ddata->d_buf = xmalloc(shdr.sh_size);
589 bcopy(sdata->d_buf, ddata->d_buf, shdr.sh_size);
590
591 for (i = 0; i < nsym; i++) {
592 GElf_Sym sym;
593 short newscn;
594
595 (void) gelf_getsym(ddata, i, &sym);
596
597 if (sym.st_shndx >= SHN_LORESERVE)
598 continue;
599
600 if ((newscn = secxlate[sym.st_shndx]) !=
601 sym.st_shndx) {
602 sym.st_shndx =
603 (newscn == -1 ? 1 : newscn);
604
605 gelf_update_sym(ddata, i, &sym);
606 }
607 }
608 }
609
610 if (gelf_update_shdr(dscn, &shdr) == 0)
611 elfterminate(dstname, "Cannot update sect %s", sname);
612
613 new_offset = (off_t)shdr.sh_offset;
614 if (shdr.sh_type != SHT_NOBITS)
615 new_offset += shdr.sh_size;
616 }
617
618 if (symtab_idx == -1) {
619 terminate("%s: Cannot find %s section\n", srcname,
620 dynsym ? "SHT_DYNSYM" : "SHT_SYMTAB");
621 }
622
623 /* Add the ctf section */
624 dscn = elf_newscn(dst);
625 gelf_getshdr(dscn, &shdr);
626 shdr.sh_name = ctfnameoff;
627 shdr.sh_type = SHT_PROGBITS;
628 shdr.sh_size = ctfsize;
629 shdr.sh_link = symtab_idx;
630 shdr.sh_addralign = 4;
631 if (changing && sehdr.e_phnum != 0) {
632 pad = new_offset % shdr.sh_addralign;
633
634 if (pad)
635 new_offset += shdr.sh_addralign - pad;
636
637 shdr.sh_offset = new_offset;
638 new_offset += shdr.sh_size;
639 }
640
641 ddata = elf_newdata(dscn);
642 ddata->d_buf = ctfdata;
643 ddata->d_size = ctfsize;
644 ddata->d_align = shdr.sh_addralign;
645
646 gelf_update_shdr(dscn, &shdr);
647
648 /* update the section header location */
649 if (sehdr.e_phnum != 0) {
650 size_t align = gelf_fsize(dst, ELF_T_ADDR, 1, EV_CURRENT);
651 size_t r = new_offset % align;
652
653 if (r)
654 new_offset += align - r;
655
656 dehdr.e_shoff = new_offset;
657 }
658
659 /* commit to disk */
660 dehdr.e_shstrndx = secxlate[sehdr.e_shstrndx];
661 gelf_update_ehdr(dst, &dehdr);
662 if (elf_update(dst, ELF_C_WRITE) < 0)
663 elfterminate(dstname, "Cannot finalize temp file");
664
665 free(secxlate);
666 }
667
668 static caddr_t
669 make_ctf_data(tdata_t *td, Elf *elf, const char *file, size_t *lenp, int flags)
670 {
671 iiburst_t *iiburst;
672 caddr_t data;
673
674 iiburst = sort_iidescs(elf, file, td, flags & CTF_FUZZY_MATCH,
675 flags & CTF_USE_DYNSYM);
676 data = ctf_gen(iiburst, lenp, flags & CTF_COMPRESS);
677
678 iiburst_free(iiburst);
679
680 return (data);
681 }
682
683 void
684 write_ctf(tdata_t *td, const char *curname, const char *newname, int flags)
685 {
686 struct stat st;
687 Elf *elf = NULL;
688 Elf *telf = NULL;
689 caddr_t data;
690 size_t len;
691 int fd = -1;
692 int tfd = -1;
693
694 (void) elf_version(EV_CURRENT);
695 if ((fd = open(curname, O_RDONLY)) < 0 || fstat(fd, &st) < 0)
696 terminate("%s: Cannot open for re-reading", curname);
697 if ((elf = elf_begin(fd, ELF_C_READ, NULL)) == NULL)
698 elfterminate(curname, "Cannot re-read");
699
700 if ((tfd = open(newname, O_RDWR | O_CREAT | O_TRUNC, st.st_mode)) < 0)
701 terminate("Cannot open temp file %s for writing", newname);
702 if ((telf = elf_begin(tfd, ELF_C_WRITE, NULL)) == NULL)
703 elfterminate(curname, "Cannot write");
704
705 data = make_ctf_data(td, elf, curname, &len, flags);
706 write_file(elf, curname, telf, newname, data, len, flags);
707 free(data);
708
709 elf_end(telf);
710 elf_end(elf);
711 (void) close(fd);
712 (void) close(tfd);
713 }