1 // SPDX-License-Identifier: GPL-2.0-or-later
3 * elf.c - ELF access library
5 * Adapted from kpatch (https://github.com/dynup/kpatch):
10 #include <sys/types.h>
23 #define MAX_NAME_LEN 128
25 static inline u32 str_hash(const char *str)
27 return jhash(str, strlen(str), 0);
30 static inline int elf_hash_bits(void)
32 return vmlinux ? ELF_HASH_BITS : 16;
35 #define elf_hash_add(hashtable, node, key) \
36 hlist_add_head(node, &hashtable[hash_min(key, elf_hash_bits())])
38 static void elf_hash_init(struct hlist_head *table)
40 __hash_init(table, 1U << elf_hash_bits());
43 #define elf_hash_for_each_possible(name, obj, member, key) \
44 hlist_for_each_entry(obj, &name[hash_min(key, elf_hash_bits())], member)
46 static void rb_add(struct rb_root *tree, struct rb_node *node,
47 int (*cmp)(struct rb_node *, const struct rb_node *))
49 struct rb_node **link = &tree->rb_node;
50 struct rb_node *parent = NULL;
54 if (cmp(node, parent) < 0)
55 link = &parent->rb_left;
57 link = &parent->rb_right;
60 rb_link_node(node, parent, link);
61 rb_insert_color(node, tree);
64 static struct rb_node *rb_find_first(const struct rb_root *tree, const void *key,
65 int (*cmp)(const void *key, const struct rb_node *))
67 struct rb_node *node = tree->rb_node;
68 struct rb_node *match = NULL;
71 int c = cmp(key, node);
77 node = node->rb_right;
84 static struct rb_node *rb_next_match(struct rb_node *node, const void *key,
85 int (*cmp)(const void *key, const struct rb_node *))
88 if (node && cmp(key, node))
93 #define rb_for_each(tree, node, key, cmp) \
94 for ((node) = rb_find_first((tree), (key), (cmp)); \
95 (node); (node) = rb_next_match((node), (key), (cmp)))
97 static int symbol_to_offset(struct rb_node *a, const struct rb_node *b)
99 struct symbol *sa = rb_entry(a, struct symbol, node);
100 struct symbol *sb = rb_entry(b, struct symbol, node);
102 if (sa->offset < sb->offset)
104 if (sa->offset > sb->offset)
107 if (sa->len < sb->len)
109 if (sa->len > sb->len)
117 static int symbol_by_offset(const void *key, const struct rb_node *node)
119 const struct symbol *s = rb_entry(node, struct symbol, node);
120 const unsigned long *o = key;
124 if (*o >= s->offset + s->len)
130 struct section *find_section_by_name(const struct elf *elf, const char *name)
134 elf_hash_for_each_possible(elf->section_name_hash, sec, name_hash, str_hash(name))
135 if (!strcmp(sec->name, name))
141 static struct section *find_section_by_index(struct elf *elf,
146 elf_hash_for_each_possible(elf->section_hash, sec, hash, idx)
153 static struct symbol *find_symbol_by_index(struct elf *elf, unsigned int idx)
157 elf_hash_for_each_possible(elf->symbol_hash, sym, hash, idx)
164 struct symbol *find_symbol_by_offset(struct section *sec, unsigned long offset)
166 struct rb_node *node;
168 rb_for_each(&sec->symbol_tree, node, &offset, symbol_by_offset) {
169 struct symbol *s = rb_entry(node, struct symbol, node);
171 if (s->offset == offset && s->type != STT_SECTION)
178 struct symbol *find_func_by_offset(struct section *sec, unsigned long offset)
180 struct rb_node *node;
182 rb_for_each(&sec->symbol_tree, node, &offset, symbol_by_offset) {
183 struct symbol *s = rb_entry(node, struct symbol, node);
185 if (s->offset == offset && s->type == STT_FUNC)
192 struct symbol *find_symbol_containing(const struct section *sec, unsigned long offset)
194 struct rb_node *node;
196 rb_for_each(&sec->symbol_tree, node, &offset, symbol_by_offset) {
197 struct symbol *s = rb_entry(node, struct symbol, node);
199 if (s->type != STT_SECTION)
206 struct symbol *find_func_containing(struct section *sec, unsigned long offset)
208 struct rb_node *node;
210 rb_for_each(&sec->symbol_tree, node, &offset, symbol_by_offset) {
211 struct symbol *s = rb_entry(node, struct symbol, node);
213 if (s->type == STT_FUNC)
220 struct symbol *find_symbol_by_name(const struct elf *elf, const char *name)
224 elf_hash_for_each_possible(elf->symbol_name_hash, sym, name_hash, str_hash(name))
225 if (!strcmp(sym->name, name))
231 struct reloc *find_reloc_by_dest_range(const struct elf *elf, struct section *sec,
232 unsigned long offset, unsigned int len)
234 struct reloc *reloc, *r = NULL;
242 for_offset_range(o, offset, offset + len) {
243 elf_hash_for_each_possible(elf->reloc_hash, reloc, hash,
244 sec_offset_hash(sec, o)) {
245 if (reloc->sec != sec)
248 if (reloc->offset >= offset && reloc->offset < offset + len) {
249 if (!r || reloc->offset < r->offset)
260 struct reloc *find_reloc_by_dest(const struct elf *elf, struct section *sec, unsigned long offset)
262 return find_reloc_by_dest_range(elf, sec, offset, 1);
265 static int read_sections(struct elf *elf)
269 size_t shstrndx, sections_nr;
272 if (elf_getshdrnum(elf->elf, §ions_nr)) {
273 WARN_ELF("elf_getshdrnum");
277 if (elf_getshdrstrndx(elf->elf, &shstrndx)) {
278 WARN_ELF("elf_getshdrstrndx");
282 for (i = 0; i < sections_nr; i++) {
283 sec = malloc(sizeof(*sec));
288 memset(sec, 0, sizeof(*sec));
290 INIT_LIST_HEAD(&sec->symbol_list);
291 INIT_LIST_HEAD(&sec->reloc_list);
293 s = elf_getscn(elf->elf, i);
295 WARN_ELF("elf_getscn");
299 sec->idx = elf_ndxscn(s);
301 if (!gelf_getshdr(s, &sec->sh)) {
302 WARN_ELF("gelf_getshdr");
306 sec->name = elf_strptr(elf->elf, shstrndx, sec->sh.sh_name);
308 WARN_ELF("elf_strptr");
312 if (sec->sh.sh_size != 0) {
313 sec->data = elf_getdata(s, NULL);
315 WARN_ELF("elf_getdata");
318 if (sec->data->d_off != 0 ||
319 sec->data->d_size != sec->sh.sh_size) {
320 WARN("unexpected data attributes for %s",
325 sec->len = sec->sh.sh_size;
327 list_add_tail(&sec->list, &elf->sections);
328 elf_hash_add(elf->section_hash, &sec->hash, sec->idx);
329 elf_hash_add(elf->section_name_hash, &sec->name_hash, str_hash(sec->name));
333 printf("nr_sections: %lu\n", (unsigned long)sections_nr);
335 /* sanity check, one more call to elf_nextscn() should return NULL */
336 if (elf_nextscn(elf->elf, s)) {
337 WARN("section entry mismatch");
344 static int read_symbols(struct elf *elf)
346 struct section *symtab, *symtab_shndx, *sec;
347 struct symbol *sym, *pfunc;
348 struct list_head *entry;
349 struct rb_node *pnode;
352 Elf_Data *shndx_data = NULL;
355 symtab = find_section_by_name(elf, ".symtab");
357 WARN("missing symbol table");
361 symtab_shndx = find_section_by_name(elf, ".symtab_shndx");
363 shndx_data = symtab_shndx->data;
365 symbols_nr = symtab->sh.sh_size / symtab->sh.sh_entsize;
367 for (i = 0; i < symbols_nr; i++) {
368 sym = malloc(sizeof(*sym));
373 memset(sym, 0, sizeof(*sym));
378 if (!gelf_getsymshndx(symtab->data, shndx_data, i, &sym->sym,
380 WARN_ELF("gelf_getsymshndx");
384 sym->name = elf_strptr(elf->elf, symtab->sh.sh_link,
387 WARN_ELF("elf_strptr");
391 sym->type = GELF_ST_TYPE(sym->sym.st_info);
392 sym->bind = GELF_ST_BIND(sym->sym.st_info);
394 if ((sym->sym.st_shndx > SHN_UNDEF &&
395 sym->sym.st_shndx < SHN_LORESERVE) ||
396 (shndx_data && sym->sym.st_shndx == SHN_XINDEX)) {
397 if (sym->sym.st_shndx != SHN_XINDEX)
398 shndx = sym->sym.st_shndx;
400 sym->sec = find_section_by_index(elf, shndx);
402 WARN("couldn't find section for symbol %s",
406 if (sym->type == STT_SECTION) {
407 sym->name = sym->sec->name;
411 sym->sec = find_section_by_index(elf, 0);
413 sym->offset = sym->sym.st_value;
414 sym->len = sym->sym.st_size;
416 rb_add(&sym->sec->symbol_tree, &sym->node, symbol_to_offset);
417 pnode = rb_prev(&sym->node);
419 entry = &rb_entry(pnode, struct symbol, node)->list;
421 entry = &sym->sec->symbol_list;
422 list_add(&sym->list, entry);
423 elf_hash_add(elf->symbol_hash, &sym->hash, sym->idx);
424 elf_hash_add(elf->symbol_name_hash, &sym->name_hash, str_hash(sym->name));
428 printf("nr_symbols: %lu\n", (unsigned long)symbols_nr);
430 /* Create parent/child links for any cold subfunctions */
431 list_for_each_entry(sec, &elf->sections, list) {
432 list_for_each_entry(sym, &sec->symbol_list, list) {
433 char pname[MAX_NAME_LEN + 1];
435 if (sym->type != STT_FUNC)
438 if (sym->pfunc == NULL)
441 if (sym->cfunc == NULL)
444 coldstr = strstr(sym->name, ".cold");
448 pnamelen = coldstr - sym->name;
449 if (pnamelen > MAX_NAME_LEN) {
450 WARN("%s(): parent function name exceeds maximum length of %d characters",
451 sym->name, MAX_NAME_LEN);
455 strncpy(pname, sym->name, pnamelen);
456 pname[pnamelen] = '\0';
457 pfunc = find_symbol_by_name(elf, pname);
460 WARN("%s(): can't find parent function",
469 * Unfortunately, -fnoreorder-functions puts the child
470 * inside the parent. Remove the overlap so we can
471 * have sane assumptions.
473 * Note that pfunc->len now no longer matches
474 * pfunc->sym.st_size.
476 if (sym->sec == pfunc->sec &&
477 sym->offset >= pfunc->offset &&
478 sym->offset + sym->len == pfunc->offset + pfunc->len) {
479 pfunc->len -= sym->len;
491 void elf_add_reloc(struct elf *elf, struct reloc *reloc)
493 struct section *sec = reloc->sec;
495 list_add_tail(&reloc->list, &sec->reloc_list);
496 elf_hash_add(elf->reloc_hash, &reloc->hash, reloc_hash(reloc));
499 static int read_rel_reloc(struct section *sec, int i, struct reloc *reloc, unsigned int *symndx)
501 if (!gelf_getrel(sec->data, i, &reloc->rel)) {
502 WARN_ELF("gelf_getrel");
505 reloc->type = GELF_R_TYPE(reloc->rel.r_info);
507 reloc->offset = reloc->rel.r_offset;
508 *symndx = GELF_R_SYM(reloc->rel.r_info);
512 static int read_rela_reloc(struct section *sec, int i, struct reloc *reloc, unsigned int *symndx)
514 if (!gelf_getrela(sec->data, i, &reloc->rela)) {
515 WARN_ELF("gelf_getrela");
518 reloc->type = GELF_R_TYPE(reloc->rela.r_info);
519 reloc->addend = reloc->rela.r_addend;
520 reloc->offset = reloc->rela.r_offset;
521 *symndx = GELF_R_SYM(reloc->rela.r_info);
525 static int read_relocs(struct elf *elf)
531 unsigned long nr_reloc, max_reloc = 0, tot_reloc = 0;
533 list_for_each_entry(sec, &elf->sections, list) {
534 if ((sec->sh.sh_type != SHT_RELA) &&
535 (sec->sh.sh_type != SHT_REL))
538 sec->base = find_section_by_index(elf, sec->sh.sh_info);
540 WARN("can't find base section for reloc section %s",
545 sec->base->reloc = sec;
548 for (i = 0; i < sec->sh.sh_size / sec->sh.sh_entsize; i++) {
549 reloc = malloc(sizeof(*reloc));
554 memset(reloc, 0, sizeof(*reloc));
555 switch (sec->sh.sh_type) {
557 if (read_rel_reloc(sec, i, reloc, &symndx))
561 if (read_rela_reloc(sec, i, reloc, &symndx))
569 reloc->sym = find_symbol_by_index(elf, symndx);
571 WARN("can't find reloc entry symbol %d for %s",
576 elf_add_reloc(elf, reloc);
579 max_reloc = max(max_reloc, nr_reloc);
580 tot_reloc += nr_reloc;
584 printf("max_reloc: %lu\n", max_reloc);
585 printf("tot_reloc: %lu\n", tot_reloc);
591 struct elf *elf_open_read(const char *name, int flags)
596 elf_version(EV_CURRENT);
598 elf = malloc(sizeof(*elf));
603 memset(elf, 0, offsetof(struct elf, sections));
605 INIT_LIST_HEAD(&elf->sections);
607 elf_hash_init(elf->symbol_hash);
608 elf_hash_init(elf->symbol_name_hash);
609 elf_hash_init(elf->section_hash);
610 elf_hash_init(elf->section_name_hash);
611 elf_hash_init(elf->reloc_hash);
613 elf->fd = open(name, flags);
615 fprintf(stderr, "objtool: Can't open '%s': %s\n",
616 name, strerror(errno));
620 if ((flags & O_ACCMODE) == O_RDONLY)
621 cmd = ELF_C_READ_MMAP;
622 else if ((flags & O_ACCMODE) == O_RDWR)
627 elf->elf = elf_begin(elf->fd, cmd, NULL);
629 WARN_ELF("elf_begin");
633 if (!gelf_getehdr(elf->elf, &elf->ehdr)) {
634 WARN_ELF("gelf_getehdr");
638 if (read_sections(elf))
641 if (read_symbols(elf))
644 if (read_relocs(elf))
654 struct section *elf_create_section(struct elf *elf, const char *name,
655 unsigned int sh_flags, size_t entsize, int nr)
657 struct section *sec, *shstrtab;
658 size_t size = entsize * nr;
662 sec = malloc(sizeof(*sec));
667 memset(sec, 0, sizeof(*sec));
669 INIT_LIST_HEAD(&sec->symbol_list);
670 INIT_LIST_HEAD(&sec->reloc_list);
672 s = elf_newscn(elf->elf);
674 WARN_ELF("elf_newscn");
678 sec->name = strdup(name);
684 sec->idx = elf_ndxscn(s);
688 sec->data = elf_newdata(s);
690 WARN_ELF("elf_newdata");
694 sec->data->d_size = size;
695 sec->data->d_align = 1;
698 sec->data->d_buf = malloc(size);
699 if (!sec->data->d_buf) {
703 memset(sec->data->d_buf, 0, size);
706 if (!gelf_getshdr(s, &sec->sh)) {
707 WARN_ELF("gelf_getshdr");
711 sec->sh.sh_size = size;
712 sec->sh.sh_entsize = entsize;
713 sec->sh.sh_type = SHT_PROGBITS;
714 sec->sh.sh_addralign = 1;
715 sec->sh.sh_flags = SHF_ALLOC | sh_flags;
718 /* Add section name to .shstrtab (or .strtab for Clang) */
719 shstrtab = find_section_by_name(elf, ".shstrtab");
721 shstrtab = find_section_by_name(elf, ".strtab");
723 WARN("can't find .shstrtab or .strtab section");
727 s = elf_getscn(elf->elf, shstrtab->idx);
729 WARN_ELF("elf_getscn");
733 data = elf_newdata(s);
735 WARN_ELF("elf_newdata");
739 data->d_buf = sec->name;
740 data->d_size = strlen(name) + 1;
743 sec->sh.sh_name = shstrtab->len;
745 shstrtab->len += strlen(name) + 1;
746 shstrtab->changed = true;
748 list_add_tail(&sec->list, &elf->sections);
749 elf_hash_add(elf->section_hash, &sec->hash, sec->idx);
750 elf_hash_add(elf->section_name_hash, &sec->name_hash, str_hash(sec->name));
757 static struct section *elf_create_rel_reloc_section(struct elf *elf, struct section *base)
762 relocname = malloc(strlen(base->name) + strlen(".rel") + 1);
767 strcpy(relocname, ".rel");
768 strcat(relocname, base->name);
770 sec = elf_create_section(elf, relocname, 0, sizeof(GElf_Rel), 0);
778 sec->sh.sh_type = SHT_REL;
779 sec->sh.sh_addralign = 8;
780 sec->sh.sh_link = find_section_by_name(elf, ".symtab")->idx;
781 sec->sh.sh_info = base->idx;
782 sec->sh.sh_flags = SHF_INFO_LINK;
787 static struct section *elf_create_rela_reloc_section(struct elf *elf, struct section *base)
792 relocname = malloc(strlen(base->name) + strlen(".rela") + 1);
797 strcpy(relocname, ".rela");
798 strcat(relocname, base->name);
800 sec = elf_create_section(elf, relocname, 0, sizeof(GElf_Rela), 0);
808 sec->sh.sh_type = SHT_RELA;
809 sec->sh.sh_addralign = 8;
810 sec->sh.sh_link = find_section_by_name(elf, ".symtab")->idx;
811 sec->sh.sh_info = base->idx;
812 sec->sh.sh_flags = SHF_INFO_LINK;
817 struct section *elf_create_reloc_section(struct elf *elf,
818 struct section *base,
822 case SHT_REL: return elf_create_rel_reloc_section(elf, base);
823 case SHT_RELA: return elf_create_rela_reloc_section(elf, base);
824 default: return NULL;
828 static int elf_rebuild_rel_reloc_section(struct section *sec, int nr)
834 /* Allocate a buffer for relocations */
835 size = nr * sizeof(*relocs);
836 relocs = malloc(size);
842 sec->data->d_buf = relocs;
843 sec->data->d_size = size;
845 sec->sh.sh_size = size;
848 list_for_each_entry(reloc, &sec->reloc_list, list) {
849 relocs[idx].r_offset = reloc->offset;
850 relocs[idx].r_info = GELF_R_INFO(reloc->sym->idx, reloc->type);
857 static int elf_rebuild_rela_reloc_section(struct section *sec, int nr)
863 /* Allocate a buffer for relocations with addends */
864 size = nr * sizeof(*relocs);
865 relocs = malloc(size);
871 sec->data->d_buf = relocs;
872 sec->data->d_size = size;
874 sec->sh.sh_size = size;
877 list_for_each_entry(reloc, &sec->reloc_list, list) {
878 relocs[idx].r_offset = reloc->offset;
879 relocs[idx].r_addend = reloc->addend;
880 relocs[idx].r_info = GELF_R_INFO(reloc->sym->idx, reloc->type);
887 int elf_rebuild_reloc_section(struct elf *elf, struct section *sec)
896 list_for_each_entry(reloc, &sec->reloc_list, list)
899 switch (sec->sh.sh_type) {
900 case SHT_REL: return elf_rebuild_rel_reloc_section(sec, nr);
901 case SHT_RELA: return elf_rebuild_rela_reloc_section(sec, nr);
906 int elf_write_insn(struct elf *elf, struct section *sec,
907 unsigned long offset, unsigned int len,
910 Elf_Data *data = sec->data;
912 if (data->d_type != ELF_T_BYTE || data->d_off) {
913 WARN("write to unexpected data for section: %s", sec->name);
917 memcpy(data->d_buf + offset, insn, len);
918 elf_flagdata(data, ELF_C_SET, ELF_F_DIRTY);
925 int elf_write_reloc(struct elf *elf, struct reloc *reloc)
927 struct section *sec = reloc->sec;
929 if (sec->sh.sh_type == SHT_REL) {
930 reloc->rel.r_info = GELF_R_INFO(reloc->sym->idx, reloc->type);
931 reloc->rel.r_offset = reloc->offset;
933 if (!gelf_update_rel(sec->data, reloc->idx, &reloc->rel)) {
934 WARN_ELF("gelf_update_rel");
938 reloc->rela.r_info = GELF_R_INFO(reloc->sym->idx, reloc->type);
939 reloc->rela.r_addend = reloc->addend;
940 reloc->rela.r_offset = reloc->offset;
942 if (!gelf_update_rela(sec->data, reloc->idx, &reloc->rela)) {
943 WARN_ELF("gelf_update_rela");
953 int elf_write(struct elf *elf)
958 /* Update section headers for changed sections: */
959 list_for_each_entry(sec, &elf->sections, list) {
961 s = elf_getscn(elf->elf, sec->idx);
963 WARN_ELF("elf_getscn");
966 if (!gelf_update_shdr(s, &sec->sh)) {
967 WARN_ELF("gelf_update_shdr");
971 sec->changed = false;
975 /* Make sure the new section header entries get updated properly. */
976 elf_flagelf(elf->elf, ELF_C_SET, ELF_F_DIRTY);
978 /* Write all changes to the file. */
979 if (elf_update(elf->elf, ELF_C_WRITE) < 0) {
980 WARN_ELF("elf_update");
984 elf->changed = false;
989 void elf_close(struct elf *elf)
991 struct section *sec, *tmpsec;
992 struct symbol *sym, *tmpsym;
993 struct reloc *reloc, *tmpreloc;
1001 list_for_each_entry_safe(sec, tmpsec, &elf->sections, list) {
1002 list_for_each_entry_safe(sym, tmpsym, &sec->symbol_list, list) {
1003 list_del(&sym->list);
1004 hash_del(&sym->hash);
1007 list_for_each_entry_safe(reloc, tmpreloc, &sec->reloc_list, list) {
1008 list_del(&reloc->list);
1009 hash_del(&reloc->hash);
1012 list_del(&sec->list);