tccelf.c

一个GNU的C语言编译器源代码,源代码过万行.

    Section *s;
    long end_offset;
    char sym_start[1024];
    char sym_end[1024];
    
    snprintf(sym_start, sizeof(sym_start), "__%s_start", section_name + 1);
    snprintf(sym_end, sizeof(sym_end), "__%s_end", section_name + 1);

    s = find_section(s1, section_name);
    if (!s) {
        end_offset = 0;
        s = data_section;
    } else {
        end_offset = s->data_offset;
    }

    add_elf_sym(symtab_section, 
                0, 0,
                ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                s->sh_num, sym_start);
    add_elf_sym(symtab_section, 
                end_offset, 0,
                ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                s->sh_num, sym_end);
}

/* add tcc runtime libraries */
static void tcc_add_runtime(TCCState *s1)
{
    char buf[1024];
    int i;
    Section *s;

    if (!s1->nostdlib) {
        snprintf(buf, sizeof(buf), "%s/%s", tcc_lib_path, "libtcc1.a");
        tcc_add_file(s1, buf);
    }
#ifdef CONFIG_TCC_BCHECK
    if (do_bounds_check) {
        unsigned long *ptr;
        Section *init_section;
        unsigned char *pinit;
        int sym_index;

        /* XXX: add an object file to do that */
        ptr = section_ptr_add(bounds_section, sizeof(unsigned long));
        *ptr = 0;
        add_elf_sym(symtab_section, 0, 0, 
                    ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                    bounds_section->sh_num, "__bounds_start");
        /* add bound check code */
        snprintf(buf, sizeof(buf), "%s/%s", tcc_lib_path, "bcheck.o");
        tcc_add_file(s1, buf);
#ifdef TCC_TARGET_I386
        if (s1->output_type != TCC_OUTPUT_MEMORY) {
            /* add 'call __bound_init()' in .init section */
            init_section = find_section(s1, ".init");
            pinit = section_ptr_add(init_section, 5);
            pinit[0] = 0xe8;
            put32(pinit + 1, -4);
            sym_index = find_elf_sym(symtab_section, "__bound_init");
            put_elf_reloc(symtab_section, init_section, 
                          init_section->data_offset - 4, R_386_PC32, sym_index);
        }
#endif
    }
#endif
    /* add libc */
    if (!s1->nostdlib) {
        tcc_add_library(s1, "c");
    }
    /* add crt end if not memory output */
    if (s1->output_type != TCC_OUTPUT_MEMORY && !s1->nostdlib) {
        tcc_add_file(s1, CONFIG_TCC_CRT_PREFIX "/crtn.o");
    }
    /* add various standard linker symbols */
    add_elf_sym(symtab_section, 
                text_section->data_offset, 0,
                ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                text_section->sh_num, "_etext");
    add_elf_sym(symtab_section, 
                data_section->data_offset, 0,
                ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                data_section->sh_num, "_edata");
    add_elf_sym(symtab_section, 
                bss_section->data_offset, 0,
                ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                bss_section->sh_num, "_end");
    /* horrible new standard ldscript defines */
    add_init_array_defines(s1, ".preinit_array");
    add_init_array_defines(s1, ".init_array");
    add_init_array_defines(s1, ".fini_array");
    
    /* add start and stop symbols for sections whose name can be
       expressed in C */
    for(i = 1; i < s1->nb_sections; i++) {
        s = s1->sections[i];
        if (s->sh_type == SHT_PROGBITS &&
            (s->sh_flags & SHF_ALLOC)) {
            const char *p;
            int ch;

            /* check if section name can be expressed in C */
            p = s->name;
            for(;;) {
                ch = *p;
                if (!ch)
                    break;
                if (!isid(ch) && !isnum(ch))
                    goto next_sec;
                p++;
            }
            snprintf(buf, sizeof(buf), "__start_%s", s->name);
            add_elf_sym(symtab_section, 
                        0, 0,
                        ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                        s->sh_num, buf);
            snprintf(buf, sizeof(buf), "__stop_%s", s->name);
            add_elf_sym(symtab_section,
                        s->data_offset, 0,
                        ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE),
                        s->sh_num, buf);
        }
    next_sec: ;
    }
}

/* name of ELF interpreter */
#ifdef __FreeBSD__
static char elf_interp[] = "/usr/libexec/ld-elf.so.1";
#else
static char elf_interp[] = "/lib/ld-linux.so.2";
#endif

#define ELF_START_ADDR 0x08048000
#define ELF_PAGE_SIZE  0x1000

/* output an ELF file */
/* XXX: suppress unneeded sections */
int tcc_output_file(TCCState *s1, const char *filename)
{
    Elf32_Ehdr ehdr;
    FILE *f;
    int fd, mode, ret;
    int *section_order;
    int shnum, i, phnum, file_offset, offset, size, j, tmp, sh_order_index, k;
    unsigned long addr;
    Section *strsec, *s;
    Elf32_Shdr shdr, *sh;
    Elf32_Phdr *phdr, *ph;
    Section *interp, *dynamic, *dynstr;
    unsigned long saved_dynamic_data_offset;
    Elf32_Sym *sym;
    int type, file_type;
    unsigned long rel_addr, rel_size;
    
    file_type = s1->output_type;
    s1->nb_errors = 0;

    if (file_type != TCC_OUTPUT_OBJ)
        tcc_add_runtime(s1);

    phdr = NULL;
    section_order = NULL;
    interp = NULL;
    dynamic = NULL;
    dynstr = NULL; /* avoid warning */
    saved_dynamic_data_offset = 0; /* avoid warning */
    
    if (file_type != TCC_OUTPUT_OBJ) {

        relocate_common_syms();

        if (!s1->static_link) {
            const char *name;
            int sym_index, index;
            Elf32_Sym *esym, *sym_end;
            
            if (file_type == TCC_OUTPUT_EXE) {
                char *ptr;
                /* add interpreter section only if executable */
                interp = new_section(s1, ".interp", SHT_PROGBITS, SHF_ALLOC);
                interp->sh_addralign = 1;
                ptr = section_ptr_add(interp, sizeof(elf_interp));
                strcpy(ptr, elf_interp);
            }
        
            /* add dynamic symbol table */
            s1->dynsym = new_symtab(s1, ".dynsym", SHT_DYNSYM, SHF_ALLOC,
                                    ".dynstr", 
                                    ".hash", SHF_ALLOC);
            dynstr = s1->dynsym->link;
            
            /* add dynamic section */
            dynamic = new_section(s1, ".dynamic", SHT_DYNAMIC, 
                                  SHF_ALLOC | SHF_WRITE);
            dynamic->link = dynstr;
            dynamic->sh_entsize = sizeof(Elf32_Dyn);
        
            /* add PLT */
            s1->plt = new_section(s1, ".plt", SHT_PROGBITS, 
                                  SHF_ALLOC | SHF_EXECINSTR);
            s1->plt->sh_entsize = 4;

            build_got(s1);

            /* scan for undefined symbols and see if they are in the
               dynamic symbols. If a symbol STT_FUNC is found, then we
               add it in the PLT. If a symbol STT_OBJECT is found, we
               add it in the .bss section with a suitable relocation */
            sym_end = (Elf32_Sym *)(symtab_section->data + 
                                    symtab_section->data_offset);
            if (file_type == TCC_OUTPUT_EXE) {
                for(sym = (Elf32_Sym *)symtab_section->data + 1; 
                    sym < sym_end;
                    sym++) {
                    if (sym->st_shndx == SHN_UNDEF) {
                        name = symtab_section->link->data + sym->st_name;
                        sym_index = find_elf_sym(s1->dynsymtab_section, name);
                        if (sym_index) {
                            esym = &((Elf32_Sym *)s1->dynsymtab_section->data)[sym_index];
                            type = ELF32_ST_TYPE(esym->st_info);
                            if (type == STT_FUNC) {
                                put_got_entry(s1, R_386_JMP_SLOT, esym->st_size, 
                                              esym->st_info, 
                                              sym - (Elf32_Sym *)symtab_section->data);
                            } else if (type == STT_OBJECT) {
                                unsigned long offset;
                                offset = bss_section->data_offset;
                                /* XXX: which alignment ? */
                                offset = (offset + 16 - 1) & -16;
                                index = put_elf_sym(s1->dynsym, offset, esym->st_size, 
                                                    esym->st_info, 0, 
                                                    bss_section->sh_num, name);
                                put_elf_reloc(s1->dynsym, bss_section, 
                                              offset, R_386_COPY, index);
                                offset += esym->st_size;
                                bss_section->data_offset = offset;
                            }
                        } else {
                                /* STB_WEAK undefined symbols are accepted */
                                /* XXX: _fp_hw seems to be part of the ABI, so we ignore
                                   it */
                            if (ELF32_ST_BIND(sym->st_info) == STB_WEAK ||
                                !strcmp(name, "_fp_hw")) {
                            } else {
                                error_noabort("undefined symbol '%s'", name);
                            }
                        }
                    } else if (s1->rdynamic && 
                               ELF32_ST_BIND(sym->st_info) != STB_LOCAL) {
                        /* if -rdynamic option, then export all non
                           local symbols */
                        name = symtab_section->link->data + sym->st_name;
                        put_elf_sym(s1->dynsym, sym->st_value, sym->st_size, 
                                    sym->st_info, 0, 
                                    sym->st_shndx, name);
                    }
                }
            
                if (s1->nb_errors)
                    goto fail;

                /* now look at unresolved dynamic symbols and export
                   corresponding symbol */
                sym_end = (Elf32_Sym *)(s1->dynsymtab_section->data + 
                                        s1->dynsymtab_section->data_offset);
                for(esym = (Elf32_Sym *)s1->dynsymtab_section->data + 1; 
                    esym < sym_end;
                    esym++) {
                    if (esym->st_shndx == SHN_UNDEF) {
                        name = s1->dynsymtab_section->link->data + esym->st_name;
                        sym_index = find_elf_sym(symtab_section, name);
                        if (sym_index) {
                            /* XXX: avoid adding a symbol if already
                               present because of -rdynamic ? */
                            sym = &((Elf32_Sym *)symtab_section->data)[sym_index];
                            put_elf_sym(s1->dynsym, sym->st_value, sym->st_size, 
                                        sym->st_info, 0, 
                                        sym->st_shndx, name);
                        } else {
                            if (ELF32_ST_BIND(esym->st_info) == STB_WEAK) {
                                /* weak symbols can stay undefined */
                            } else {
                                warning("undefined dynamic symbol '%s'", name);
                            }
                        }
                    }
                }
            } else {
                int nb_syms;
                /* shared library case : we simply export all the global symbols */
                nb_syms = symtab_section->data_offset / sizeof(Elf32_Sym);
                s1->symtab_to_dynsym = tcc_mallocz(sizeof(int) * nb_syms);
                for(sym = (Elf32_Sym *)symtab_section->data + 1; 
                    sym < sym_end;
                    sym++) {
                    if (ELF32_ST_BIND(sym->st_info) != STB_LOCAL) {
                        name = symtab_section->link->data + sym->st_name;
                        index = put_elf_sym(s1->dynsym, sym->st_value, sym->st_size, 
                                            sym->st_info, 0, 
                                            sym->st_shndx, name);
                        s1->symtab_to_dynsym[sym - 
                                            (Elf32_Sym *)symtab_section->data] = 
                            index;
                    }
                }
            }

            build_got_entries(s1);
        
            /* add a list of needed dlls */
            for(i = 0; i < s1->nb_loaded_dlls; i++) {
                DLLReference *dllref = s1->loaded_dlls[i];
                if (dllref->level == 0)
                    put_dt(dynamic, DT_NEEDED, put_elf_str(dynstr, dllref->name));
            }
            /* XXX: currently, since we do not handle PIC code, we
               must relocate the readonly segments */
            if (file_type == TCC_OUTPUT_DLL)
                put_dt(dynamic, DT_TEXTREL, 0);

            /* add necessary space for other entries */
            saved_dynamic_data_offset = dynamic->data_offset;
            dynamic->data_offset += 8 * 9;
        } else {
            /* still need to build got entries in case of static link */
            build_got_entries(s1);
        }
    }

    memset(&ehdr, 0, sizeof(ehdr));

    /* we add a section for symbols */
    strsec = new_section(s1, ".shstrtab", SHT_STRTAB, 0);
    put_elf_str(strsec, "");
    
    /* compute number of sections */
    shnum = s1->nb_sections;

    /* this array is used to reorder sections in the output file */
    section_order = tcc_malloc(sizeof(int) * shnum);
    section_order[0] = 0;
    sh_order_index = 1;
    
    /* compute number of program headers */
    switch(file_type) {
    default:
    case TCC_OUTPUT_OBJ:
        phnum = 0;
        break;
    case TCC_OUTPUT_EXE:
        if (!s1->static_link)
            phnum = 4;
        else
            phnum = 2;
        break;
    case TCC_OUTPUT_DLL:
        phnum = 3;
        break;
    }

    /* allocate strings for section names and decide if an unallocated
       section should be output */
    /* NOTE: the strsec section comes last, so its size is also
       correct ! */
    for(i = 1; i < s1->nb_sections; i++) {
        s = s1->sections[i];
        s->sh_name = put_elf_str(strsec, s->name);
        /* when generating a DLL, we include relocations but we may
           patch them */
        if (file_type == TCC_OUTPUT_DLL && 
            s->sh_type == SHT_REL && 
            !(s->sh_flags & SHF_ALLOC)) {
            prepare_dynamic_rel(s1, s);
        } else if (do_debug || 
            file_type == TCC_OUTPUT_OBJ || 
            (s->sh_flags & SHF_ALLOC) ||
            i == (s1->nb_sections - 1)) {
            /* we output all sections if debug or object file */
            s->sh_size = s->data_offset;
        }
    }

    /* allocate program segment headers */
    phdr = tcc_mallocz(phnum * sizeof(Elf32_Phdr));
        
    file_offset = sizeof(Elf32_Ehdr) + phnum * sizeof(Elf32_Phdr);
    if (phnum > 0) {
        /* compute section to program header mapping */
        if (file_type == TCC_OUTPUT_DLL)
            addr = 0;