| /* |
| * Copyright (c) 2011-2013 ARM Limited |
| * All rights reserved |
| * |
| * The license below extends only to copyright in the software and shall |
| * not be construed as granting a license to any other intellectual |
| * property including but not limited to intellectual property relating |
| * to a hardware implementation of the functionality of the software |
| * licensed hereunder. You may use the software subject to the license |
| * terms below provided that you ensure that this notice is replicated |
| * unmodified and in its entirety in all distributions of the software, |
| * modified or unmodified, in source code or in binary form. |
| * |
| * Copyright (c) 2003-2005 The Regents of The University of Michigan |
| * All rights reserved. |
| * |
| * Redistribution and use in source and binary forms, with or without |
| * modification, are permitted provided that the following conditions are |
| * met: redistributions of source code must retain the above copyright |
| * notice, this list of conditions and the following disclaimer; |
| * redistributions in binary form must reproduce the above copyright |
| * notice, this list of conditions and the following disclaimer in the |
| * documentation and/or other materials provided with the distribution; |
| * neither the name of the copyright holders nor the names of its |
| * contributors may be used to endorse or promote products derived from |
| * this software without specific prior written permission. |
| * |
| * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
| * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
| * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR |
| * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT |
| * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, |
| * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
| * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
| * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
| * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
| * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| * |
| * Authors: Steve Reinhardt |
| * Ali Saidi |
| */ |
| |
| #include "base/loader/elf_object.hh" |
| |
| #include <fcntl.h> |
| #include <sys/mman.h> |
| #include <sys/stat.h> |
| #include <sys/types.h> |
| #include <unistd.h> |
| |
| #include <cassert> |
| #include <string> |
| |
| #include "base/bitfield.hh" |
| #include "base/loader/symtab.hh" |
| #include "base/logging.hh" |
| #include "base/trace.hh" |
| #include "debug/Loader.hh" |
| #include "gelf.h" |
| #include "sim/byteswap.hh" |
| |
| ObjectFile * |
| ElfObject::tryFile(const std::string &fname, size_t len, uint8_t *data, |
| bool skip_interp_check) |
| { |
| // check that header matches library version |
| if (elf_version(EV_CURRENT) == EV_NONE) |
| panic("wrong elf version number!"); |
| |
| // get a pointer to elf structure |
| // Check that we actually have a elf file |
| Elf *elf = elf_memory((char*)data, len); |
| assert(elf); |
| |
| GElf_Ehdr ehdr; |
| if (gelf_getehdr(elf, &ehdr) == 0) { |
| DPRINTFR(Loader, "Not ELF\n"); |
| elf_end(elf); |
| return NULL; |
| } |
| |
| // Detect the architecture |
| Arch arch = UnknownArch; |
| if (ehdr.e_machine == EM_SPARC64 || |
| (ehdr.e_machine == EM_SPARC && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS64) || |
| ehdr.e_machine == EM_SPARCV9) { |
| arch = SPARC64; |
| } else if (ehdr.e_machine == EM_SPARC32PLUS || |
| (ehdr.e_machine == EM_SPARC && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS32)) { |
| arch = SPARC32; |
| } else if (ehdr.e_machine == EM_MIPS && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS32) { |
| arch = Mips; |
| if (ehdr.e_ident[EI_DATA] != ELFDATA2LSB) { |
| fatal("The binary you're trying to load is compiled for big " |
| "endian MIPS. gem5\nonly supports little endian MIPS. " |
| "Please recompile your binary.\n"); |
| } |
| } else if (ehdr.e_machine == EM_X86_64 && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS64) { |
| arch = X86_64; |
| } else if (ehdr.e_machine == EM_386 && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS32) { |
| arch = I386; |
| } else if (ehdr.e_machine == EM_ARM && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS32) { |
| arch = bits(ehdr.e_entry, 0) ? Thumb : Arm; |
| } else if (ehdr.e_machine == EM_AARCH64 && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS64) { |
| arch = Arm64; |
| } else if (ehdr.e_machine == EM_RISCV) { |
| arch = (ehdr.e_ident[EI_CLASS] == ELFCLASS64) ? Riscv64 : Riscv32; |
| } else if (ehdr.e_machine == EM_PPC && |
| ehdr.e_ident[EI_CLASS] == ELFCLASS32) { |
| arch = Power; |
| if (ehdr.e_ident[EI_DATA] != ELFDATA2MSB) { |
| fatal("The binary you're trying to load is compiled for " |
| "little endian Power.\ngem5 only supports big " |
| "endian Power. Please recompile your binary.\n"); |
| } |
| } else if (ehdr.e_machine == EM_PPC64) { |
| fatal("The binary you're trying to load is compiled for 64-bit " |
| "Power. M5\n only supports 32-bit Power. Please " |
| "recompile your binary.\n"); |
| } else if (ehdr.e_ident[EI_CLASS] == ELFCLASS64) { |
| // Since we don't know how to check for alpha right now, we'll |
| // just assume if it wasn't something else and it's 64 bit, that's |
| // what it must be. |
| arch = Alpha; |
| } else { |
| warn("Unknown architecture: %d\n", ehdr.e_machine); |
| arch = UnknownArch; |
| } |
| |
| // Detect the operating system |
| OpSys op_sys; |
| switch (ehdr.e_ident[EI_OSABI]) { |
| case ELFOSABI_LINUX: |
| op_sys = Linux; |
| break; |
| case ELFOSABI_SOLARIS: |
| op_sys = Solaris; |
| break; |
| case ELFOSABI_TRU64: |
| op_sys = Tru64; |
| break; |
| case ELFOSABI_ARM: |
| op_sys = LinuxArmOABI; |
| break; |
| case ELFOSABI_FREEBSD: |
| op_sys = FreeBSD; |
| break; |
| default: |
| op_sys = UnknownOpSys; |
| } |
| |
| // Take a look at the .note.ABI section. |
| // It can let us know what's what. |
| if (op_sys == UnknownOpSys) { |
| int sec_idx = 1; |
| |
| // Get the first section |
| Elf_Scn *section = elf_getscn(elf, sec_idx); |
| |
| // While there are no more sections |
| while (section && op_sys == UnknownOpSys) { |
| GElf_Shdr shdr; |
| gelf_getshdr(section, &shdr); |
| |
| char *e_str = elf_strptr(elf, ehdr.e_shstrndx, shdr.sh_name); |
| if (shdr.sh_type == SHT_NOTE && |
| !strcmp(".note.ABI-tag", e_str)) { |
| // we have found a ABI note section |
| // Check the 5th 32bit word for OS 0 == linux, 1 == hurd, |
| // 2 == solaris, 3 == freebsd |
| Elf_Data *raw_data = elf_rawdata(section, NULL); |
| assert(raw_data && raw_data->d_buf); |
| |
| uint32_t raw_abi = ((uint32_t*)raw_data->d_buf)[4]; |
| bool is_le = ehdr.e_ident[EI_DATA] == ELFDATA2LSB; |
| uint32_t os_abi = is_le ? htole(raw_abi) : htobe(raw_abi); |
| |
| switch (os_abi) { |
| case 0: |
| op_sys = Linux; |
| break; |
| case 1: |
| fatal("gem5 does not support the HURD ABI.\n"); |
| case 2: |
| op_sys = Solaris; |
| break; |
| case 3: |
| op_sys = FreeBSD; |
| break; |
| } |
| } // if section found |
| |
| if (!strcmp(".SUNW_version", e_str) || |
| !strcmp(".stab.index", e_str)) |
| op_sys = Solaris; |
| |
| section = elf_getscn(elf, ++sec_idx); |
| } // while sections |
| } |
| |
| ElfObject * result = new ElfObject(fname, len, data, arch, op_sys); |
| |
| // The number of headers in the file |
| result->_programHeaderCount = ehdr.e_phnum; |
| // Record the size of each entry |
| result->_programHeaderSize = ehdr.e_phentsize; |
| result->_programHeaderTable = 0; |
| if (result->_programHeaderCount) { // If there is a program header table |
| // Figure out the virtual address of the header table in the |
| // final memory image. We use the program headers themselves |
| // to translate from a file offset to the address in the image. |
| GElf_Phdr phdr; |
| uint64_t e_phoff = ehdr.e_phoff; |
| |
| for (int i = 0; i < result->_programHeaderCount; i++) { |
| gelf_getphdr(elf, i, &phdr); |
| // Check if we've found the segment with the headers in it |
| if (phdr.p_offset <= e_phoff && |
| phdr.p_offset + phdr.p_filesz > e_phoff) { |
| result->_programHeaderTable = |
| phdr.p_paddr + (e_phoff - phdr.p_offset); |
| break; |
| } |
| } |
| } |
| |
| if (!skip_interp_check) { |
| for (int i = 0; i < ehdr.e_phnum; i++) { |
| GElf_Phdr phdr; |
| M5_VAR_USED void *check_p = gelf_getphdr(elf, i, &phdr); |
| assert(check_p != nullptr); |
| |
| if (phdr.p_type != PT_INTERP) |
| continue; |
| |
| char *interp_path = (char*)data + phdr.p_offset; |
| int fd = open(interp_path, O_RDONLY); |
| if (fd == -1) |
| fatal("Unable to open dynamic executable's interpreter.\n"); |
| |
| struct stat sb; |
| M5_VAR_USED int check_i = fstat(fd, &sb); |
| assert(check_i == 0); |
| |
| void *mm = mmap(nullptr, sb.st_size, PROT_READ, |
| MAP_PRIVATE, fd, 0); |
| assert(mm != MAP_FAILED); |
| close(fd); |
| |
| uint8_t *interp_image = (uint8_t*)mm; |
| ObjectFile *obj = tryFile(interp_path, sb.st_size, |
| interp_image, true); |
| assert(obj != nullptr); |
| result->interpreter = dynamic_cast<ElfObject*>(obj); |
| assert(result->interpreter != nullptr); |
| break; |
| } |
| } |
| |
| elf_end(elf); |
| return result; |
| } |
| |
| ElfObject::ElfObject(const std::string &_filename, size_t _len, |
| uint8_t *_data, Arch _arch, OpSys _op_sys) |
| : ObjectFile(_filename, _len, _data, _arch, _op_sys), |
| _programHeaderTable(0), _programHeaderSize(0), _programHeaderCount(0), |
| interpreter(nullptr), ldBias(0), relocate(true), |
| ldMin(std::numeric_limits<Addr>::max()), |
| ldMax(std::numeric_limits<Addr>::min()) |
| { |
| // check that header matches library version |
| if (elf_version(EV_CURRENT) == EV_NONE) |
| panic("wrong elf version number!"); |
| |
| // get a pointer to elf structure |
| Elf *elf = elf_memory((char*)fileData,len); |
| assert(elf); |
| |
| // Check that we actually have a elf file |
| GElf_Ehdr ehdr; |
| if (gelf_getehdr(elf, &ehdr) ==0) { |
| panic("Not ELF, shouldn't be here"); |
| } |
| |
| entry = ehdr.e_entry; |
| |
| // initialize segment sizes to 0 in case they're not present |
| text.size = data.size = bss.size = 0; |
| text.baseAddr = data.baseAddr = bss.baseAddr = 0; |
| |
| int sec_idx = 1; |
| |
| // The first address of some important sections. |
| Addr text_sec_start = 0; |
| Addr data_sec_start = 0; |
| Addr bss_sec_start = 0; |
| |
| // Get the first section |
| Elf_Scn *section = elf_getscn(elf, sec_idx); |
| |
| // Find the beginning of the most interesting sections. |
| while (section) { |
| GElf_Shdr shdr; |
| gelf_getshdr(section, &shdr); |
| char *sec_name = elf_strptr(elf, ehdr.e_shstrndx, shdr.sh_name); |
| |
| if (sec_name) { |
| if (!strcmp(".text", sec_name)) { |
| text_sec_start = shdr.sh_addr; |
| } else if (!strcmp(".data", sec_name)) { |
| data_sec_start = shdr.sh_addr; |
| } else if (!strcmp(".bss", sec_name)) { |
| bss_sec_start = shdr.sh_addr; |
| } |
| } else { |
| Elf_Error errorNum = (Elf_Error)elf_errno(); |
| if (errorNum != ELF_E_NONE) { |
| const char *errorMessage = elf_errmsg(errorNum); |
| fatal("Error from libelf: %s.\n", errorMessage); |
| } |
| } |
| |
| section = elf_getscn(elf, ++sec_idx); |
| } |
| |
| // Go through all the segments in the program, record them, and scrape |
| // out information about the text, data, and bss areas needed by other |
| // code. |
| for (int i = 0; i < ehdr.e_phnum; ++i) { |
| GElf_Phdr phdr; |
| if (gelf_getphdr(elf, i, &phdr) == 0) { |
| panic("gelf_getphdr failed for segment %d.", i); |
| } |
| |
| // for now we don't care about non-loadable segments |
| if (!(phdr.p_type & PT_LOAD)) |
| continue; |
| |
| ldMin = std::min(ldMin, phdr.p_vaddr); |
| ldMax = std::max(ldMax, phdr.p_vaddr + phdr.p_memsz); |
| |
| // Check to see if this segment contains the bss section. |
| if (phdr.p_paddr <= bss_sec_start && |
| phdr.p_paddr + phdr.p_memsz > bss_sec_start && |
| phdr.p_memsz - phdr.p_filesz > 0) { |
| bss.baseAddr = phdr.p_paddr + phdr.p_filesz; |
| bss.size = phdr.p_memsz - phdr.p_filesz; |
| bss.fileImage = NULL; |
| } |
| |
| // Check to see if this is the text or data segment |
| if (phdr.p_vaddr <= text_sec_start && |
| phdr.p_vaddr + phdr.p_filesz > text_sec_start) { |
| |
| // If this value is nonzero, we need to flip the relocate flag. |
| if (phdr.p_vaddr != 0) |
| relocate = false; |
| |
| text.baseAddr = phdr.p_paddr; |
| text.size = phdr.p_filesz; |
| text.fileImage = fileData + phdr.p_offset; |
| } else if (phdr.p_vaddr <= data_sec_start && |
| phdr.p_vaddr + phdr.p_filesz > data_sec_start) { |
| data.baseAddr = phdr.p_paddr; |
| data.size = phdr.p_filesz; |
| data.fileImage = fileData + phdr.p_offset; |
| } else { |
| // If it's none of the above but is loadable, |
| // load the filesize worth of data |
| Segment extra; |
| extra.baseAddr = phdr.p_paddr; |
| extra.size = phdr.p_filesz; |
| extra.fileImage = fileData + phdr.p_offset; |
| extraSegments.push_back(extra); |
| } |
| } |
| |
| // should have found at least one loadable segment |
| warn_if(text.size == 0, |
| "Empty .text segment in '%s'. ELF file corrupted?\n", |
| filename); |
| |
| DPRINTFR(Loader, "text: 0x%x %d\ndata: 0x%x %d\nbss: 0x%x %d\n", |
| text.baseAddr, text.size, data.baseAddr, data.size, |
| bss.baseAddr, bss.size); |
| |
| elf_end(elf); |
| |
| // We will actually read the sections when we need to load them |
| } |
| |
| |
| bool |
| ElfObject::loadSomeSymbols(SymbolTable *symtab, int binding, Addr mask, |
| Addr base, Addr offset) |
| { |
| if (!symtab) |
| return false; |
| |
| // check that header matches library version |
| if (elf_version(EV_CURRENT) == EV_NONE) |
| panic("wrong elf version number!"); |
| |
| // get a pointer to elf structure |
| Elf *elf = elf_memory((char*)fileData,len); |
| assert(elf != NULL); |
| |
| // Get the first section |
| int sec_idx = 1; // there is a 0 but it is nothing, go figure |
| Elf_Scn *section = elf_getscn(elf, sec_idx); |
| |
| // While there are no more sections |
| bool found = false; |
| while (section != NULL) { |
| GElf_Shdr shdr; |
| gelf_getshdr(section, &shdr); |
| |
| if (shdr.sh_type == SHT_SYMTAB) { |
| found = true; |
| Elf_Data *data = elf_getdata(section, NULL); |
| int count = shdr.sh_size / shdr.sh_entsize; |
| DPRINTF(Loader, "Found Symbol Table, %d symbols present\n", count); |
| |
| // loop through all the symbols, only loading global ones |
| for (int i = 0; i < count; ++i) { |
| GElf_Sym sym; |
| gelf_getsym(data, i, &sym); |
| if (GELF_ST_BIND(sym.st_info) == binding) { |
| char *sym_name = elf_strptr(elf, shdr.sh_link, sym.st_name); |
| if (sym_name && sym_name[0] != '$') { |
| Addr value = sym.st_value - base + offset; |
| if (symtab->insert(value & mask, sym_name)) { |
| DPRINTF(Loader, "Symbol: %-40s value %#x\n", |
| sym_name, value); |
| } |
| } |
| } |
| } |
| } |
| ++sec_idx; |
| section = elf_getscn(elf, sec_idx); |
| } |
| |
| elf_end(elf); |
| |
| return found; |
| } |
| |
| bool |
| ElfObject::loadAllSymbols(SymbolTable *symtab, Addr base, Addr offset, |
| Addr addr_mask) |
| { |
| return (loadGlobalSymbols(symtab, base, offset, addr_mask) && |
| loadLocalSymbols(symtab, base, offset, addr_mask) && |
| loadWeakSymbols(symtab, base, offset, addr_mask)); |
| } |
| |
| bool |
| ElfObject::loadGlobalSymbols(SymbolTable *symtab, Addr base, Addr offset, |
| Addr addr_mask) |
| { |
| if (interpreter) { |
| interpreter->loadSomeSymbols(symtab, STB_GLOBAL, addr_mask, |
| base, offset); |
| } |
| return loadSomeSymbols(symtab, STB_GLOBAL, addr_mask, base, offset); |
| } |
| |
| bool |
| ElfObject::loadLocalSymbols(SymbolTable *symtab, Addr base, Addr offset, |
| Addr addr_mask) |
| { |
| if (interpreter) { |
| interpreter->loadSomeSymbols(symtab, STB_LOCAL, addr_mask, |
| base, offset); |
| } |
| return loadSomeSymbols(symtab, STB_LOCAL, addr_mask, base, offset); |
| } |
| |
| bool |
| ElfObject::loadWeakSymbols(SymbolTable *symtab, Addr base, Addr offset, |
| Addr addr_mask) |
| { |
| if (interpreter) { |
| interpreter->loadSomeSymbols(symtab, STB_WEAK, addr_mask, |
| base, offset); |
| } |
| return loadSomeSymbols(symtab, STB_WEAK, addr_mask, base, offset); |
| } |
| |
| bool |
| ElfObject::loadSections(const PortProxy& mem_proxy, Addr addr_mask, |
| Addr offset) |
| { |
| if (!ObjectFile::loadSections(mem_proxy, addr_mask, offset)) |
| return false; |
| |
| for (auto seg : extraSegments) { |
| if (!loadSection(&seg, mem_proxy, addr_mask, offset)) { |
| return false; |
| } |
| } |
| |
| if (interpreter) |
| interpreter->loadSections(mem_proxy, addr_mask, offset); |
| |
| return true; |
| } |
| |
| void |
| ElfObject::getSections() |
| { |
| assert(!sectionNames.size()); |
| |
| // check that header matches library version |
| if (elf_version(EV_CURRENT) == EV_NONE) |
| panic("wrong elf version number!"); |
| |
| // get a pointer to elf structure |
| Elf *elf = elf_memory((char*)fileData,len); |
| assert(elf != NULL); |
| |
| // Check that we actually have a elf file |
| GElf_Ehdr ehdr; |
| if (gelf_getehdr(elf, &ehdr) ==0) { |
| panic("Not ELF, shouldn't be here"); |
| } |
| |
| // Get the first section |
| int sec_idx = 1; // there is a 0 but it is nothing, go figure |
| Elf_Scn *section = elf_getscn(elf, sec_idx); |
| |
| // While there are no more sections |
| while (section) { |
| GElf_Shdr shdr; |
| gelf_getshdr(section, &shdr); |
| sectionNames.insert(elf_strptr(elf, ehdr.e_shstrndx, shdr.sh_name)); |
| section = elf_getscn(elf, ++sec_idx); |
| } // while sections |
| |
| elf_end(elf); |
| } |
| |
| bool |
| ElfObject::sectionExists(std::string sec) |
| { |
| if (!sectionNames.size()) |
| getSections(); |
| |
| return sectionNames.find(sec) != sectionNames.end(); |
| } |
| |
| |
| void |
| ElfObject::updateBias(Addr bias_addr) |
| { |
| // Record the bias. |
| ldBias = bias_addr; |
| |
| // Patch the entry point with bias_addr. |
| entry += bias_addr; |
| |
| // Patch segments with the bias_addr. |
| text.baseAddr += bias_addr; |
| data.baseAddr += bias_addr; |
| bss.baseAddr += bias_addr; |
| for (auto &segment : extraSegments) |
| segment.baseAddr += bias_addr; |
| } |
