[原创] 细说So动态库的加载流程
博客：www.wireghost.cn
细说So动态库的加载流程
dlopen之内存装载
dlopen用来打开一个动态链接库，并将其装入内存。它的定义在Android源码中的路径为/bionic/linker/dlfcn.cpp，执行流程如下：
其核心代码在do_dlopen中实现，根据传入的路径或文件名去查找一个动态库，并执行该动态链接库的初始化代码。
void* dlopen(const char* filename, int flags) {
ScopedPthreadMutexLocker locker(&gDlMutex);
soinfo* result = do_dlopen(filename, flags);
if (result == NULL) {
__bionic_format_dlerror("dlopen failed", linker_get_error_buffer());
return NULL;
}
return result;
}
soinfo* do_dlopen(const char* name, int flags) {
if ((flags & ~(RTLD_NOW|RTLD_LAZY|RTLD_LOCAL|RTLD_GLOBAL)) != 0) {
DL_ERR("invalid flags to dlopen: %x", flags);
return NULL;
}
set_soinfo_pool_protection(PROT_READ | PROT_WRITE);
soinfo* si = find_library(name);
if (si != NULL) {
si->CallConstructors();
}
set_soinfo_pool_protection(PROT_READ);
return si;
}
再来看find_library这个方法，它会先在solist（已经加载的动态链接库链表）里进行查找，如果找到了就返回对应的soinfo结构体指针。否则，就调用load_library进行加载。然后，调用soinfo_link_image方法，根据soinfo结构体解析相应的Section。
static soinfo *find_loaded_library(const char *name)
{
soinfo *si;
const char *bname;
// TODO: don't use basename only for determining libraries
// http://code.google.com/p/android/issues/detail?id=6670
bname = strrchr(name, '/');
bname = bname ? bname + 1 : name;
for (si = solist; si != NULL; si = si->next) {
if (!strcmp(bname, si->name)) {
return si;
}
}
return NULL;
}
static soinfo* find_library_internal(const char* name) {
if (name == NULL) {
return somain;
}
soinfo* si = find_loaded_library(name);
if (si != NULL) {
if (si->flags & FLAG_LINKED) {
return si;
}
DL_ERR("OOPS: recursive link to \"%s\"", si->name);
return NULL;
}
TRACE("[ '%s' has not been loaded yet. Locating...]", name);
si = load_library(name);
if (si == NULL) {
return NULL;
}
// At this point we know that whatever is loaded @ base is a valid ELF
// shared library whose segments are properly mapped in.
TRACE("[ init_library base=0x%08x sz=0x%08x name='%s' ]",
si->base, si->size, si->name);
if (!soinfo_link_image(si)) {
munmap(reinterpret_cast<void*>(si->base), si->size);
soinfo_free(si);
return NULL;
}
return si;
}
static soinfo* find_library(const char* name) {
soinfo* si = find_library_internal(name);
if (si != NULL) {
si->ref_count++;
}
return si;
}
load_library调用open_library打开一个动态链接库，返回一个句柄，将其与共享库所在的路径作为参数，对ElfReader进行初始化。
ElfReader作用域中的Load函数，会执行以下操作：
- 读取并校验ELF文件头
- 读ELF程序头并映射至内存
- 将Load Segment加载进内存
- 在内存中找到程序的起始地址
``` C++
bool ElfReader::Load() {
return ReadElfHeader() &&VerifyElfHeader() && ReadProgramHeader() && ReserveAddressSpace() && LoadSegments() && FindPhdr();
}
bool ElfReader::ReadElfHeader() {
ssize_t rc = TEMP_FAILURERETRY(read(fd, &header, sizeof(header)));
if (rc < 0) {
DLERR("can't read file \"%s\": %s", name, strerror(errno));
return false;
}
if (rc != sizeof(header_)) {
DLERR("\"%s\" is too small to be an ELF executable", name);
return false;
}
return true;
}
<center><font color=#006666 size=3 face="黑体">**读ELF文件头**</font></center>
``` C++
// Loads the program header table from an ELF file into a read-only private
// anonymous mmap-ed block.
bool ElfReader::ReadProgramHeader() {
phdr_num_ = header_.e_phnum;
// Like the kernel, we only accept program header tables that
// are smaller than 64KiB.
if (phdr_num_ < 1 || phdr_num_ > 65536/sizeof(Elf32_Phdr)) {
DL_ERR("\"%s\" has invalid e_phnum: %d", name_, phdr_num_);
return false;
}
Elf32_Addr page_min = PAGE_START(header_.e_phoff); //页的起始地址
Elf32_Addr page_max = PAGE_END(header_.e_phoff + (phdr_num_ * sizeof(Elf32_Phdr))); //页的结束地址
Elf32_Addr page_offset = PAGE_OFFSET(header_.e_phoff); //程序头部在页中的偏移
phdr_size_ = page_max - page_min;
void* mmap_result = mmap(NULL, phdr_size_, PROT_READ, MAP_PRIVATE, fd_, page_min); //将程序头映射到内存
if (mmap_result == MAP_FAILED) {
DL_ERR("\"%s\" phdr mmap failed: %s", name_, strerror(errno));
return false;
}
phdr_mmap_ = mmap_result;
phdr_table_ = reinterpret_cast<Elf32_Phdr*>(reinterpret_cast<char*>(mmap_result) + page_offset); //程序头表在内存中的地址
return true;
}
<center><font color=#006666 size=3 face="黑体">读ELF程序头，并映射到内存</font></center>
// Reserve a virtual address range big enough to hold all loadable
// segments of a program header table. This is done by creating a
// private anonymous mmap() with PROT_NONE.
bool ElfReader::ReserveAddressSpace() {
Elf32_Addr min_vaddr;
load_size_ = phdr_table_get_load_size(phdr_table_, phdr_num_, &min_vaddr); //根据页对齐来计算Load段所占用的大小
if (load_size_ == 0) {
DL_ERR("\"%s\" has no loadable segments", name_);
return false;
}
uint8_t* addr = reinterpret_cast<uint8_t*>(min_vaddr);
int mmap_flags = MAP_PRIVATE | MAP_ANONYMOUS; //匿名私有
void* start = mmap(addr, load_size_, PROT_NONE, mmap_flags, -1, 0); //调用mmap为动态库分配一块内存空间
if (start == MAP_FAILED) {
DL_ERR("couldn't reserve %d bytes of address space for \"%s\"", load_size_, name_);
return false;
}
load_start_ = start;
load_bias_ = reinterpret_cast<uint8_t*>(start) - addr; //真实的加载地址与计算出来的（读ELF程序头中的p_vaddr）加载地址之差
return true;
}
<center><font color=#006666 size=3 face="黑体">调用mmap申请一块足够大的内存空间，为后面进行映射Load段的映射做准备</font></center>
// Map all loadable segments in process' address space.
// This assumes you already called phdr_table_reserve_memory to
// reserve the address space range for the library.
// TODO: assert assumption.
bool ElfReader::LoadSegments() {
for (size_t i = 0; i < phdr_num_; ++i) {
const Elf32_Phdr* phdr = &phdr_table_[i];
if (phdr->p_type != PT_LOAD) {
continue;
}
// Segment addresses in memory.
Elf32_Addr seg_start = phdr->p_vaddr + load_bias_;
Elf32_Addr seg_end = seg_start + phdr->p_memsz;
Elf32_Addr seg_page_start = PAGE_START(seg_start);
Elf32_Addr seg_page_end = PAGE_END(seg_end);
Elf32_Addr seg_file_end = seg_start + phdr->p_filesz;
// File offsets.
Elf32_Addr file_start = phdr->p_offset;
Elf32_Addr file_end = file_start + phdr->p_filesz;
Elf32_Addr file_page_start = PAGE_START(file_start);
Elf32_Addr file_length = file_end - file_page_start;
if (file_length != 0) {
void* seg_addr = mmap((void*)seg_page_start, //将Load Segment映射到内存，大小为在ELF文件中所占用的长度
file_length,
PFLAGS_TO_PROT(phdr->p_flags),
MAP_FIXED|MAP_PRIVATE,
fd_,
file_page_start);
if (seg_addr == MAP_FAILED) {
DL_ERR("couldn't map \"%s\" segment %d: %s", name_, i, strerror(errno));
return false;
}
}
// if the segment is writable, and does not end on a page boundary,
// zero-fill it until the page limit.
if ((phdr->p_flags & PF_W) != 0 && PAGE_OFFSET(seg_file_end) > 0) {
memset((void*)seg_file_end, 0, PAGE_SIZE - PAGE_OFFSET(seg_file_end)); //如果这块Segment是可写的，且在内存中的结束地址不在页的边界上，则将后面的数据都填充0
}
seg_file_end = PAGE_END(seg_file_end);
// seg_file_end is now the first page address after the file
// content. If seg_end is larger, we need to zero anything
// between them. This is done by using a private anonymous
// map for all extra pages.
if (seg_page_end > seg_file_end) {
void* zeromap = mmap((void*)seg_file_end, //如果seg_end大于它在文件中的长度，则继续为多出的那部分申请内存空间，并填充0。这里应该是主要针对bss段
seg_page_end - seg_file_end,
PFLAGS_TO_PROT(phdr->p_flags),
MAP_FIXED|MAP_ANONYMOUS|MAP_PRIVATE,
-1,
0);
if (zeromap == MAP_FAILED) {
DL_ERR("couldn't zero fill \"%s\" gap: %s", name_, strerror(errno));
return false;
}
}
}
return true;
}
<center><font color=#006666 size=3 face="黑体">将类型为Load的Segment映射到内存</font></center>
接下来，soinfo_alloc方法会为该库在共享库链表中分配一个soinfo节点，并初始化其数据结构。
static soinfo* load_library(const char* name) {
// Open the file.
int fd = open_library(name);
if (fd == -1) {
DL_ERR("library \"%s\" not found", name);
return NULL;
}
// Read the ELF header and load the segments.
ElfReader elf_reader(name, fd);
if (!elf_reader.Load()) {
return NULL;
}
const char* bname = strrchr(name, '/');
soinfo* si = soinfo_alloc(bname ? bname + 1 : name);
if (si == NULL) {
return NULL;
}
si->base = elf_reader.load_start();
si->size = elf_reader.load_size();
si->load_bias = elf_reader.load_bias();
si->flags = 0;
si->entry = 0;
si->dynamic = NULL;
si->phnum = elf_reader.phdr_count();
si->phdr = elf_reader.loaded_phdr();
return si;
}
static soinfo* soinfo_alloc(const char* name) {
if (strlen(name) >= SOINFO_NAME_LEN) {
DL_ERR("library name \"%s\" too long", name);
return NULL;
}
if (!ensure_free_list_non_empty()) {
DL_ERR("out of memory when loading \"%s\"", name);
return NULL;
}
// Take the head element off the free list.
soinfo* si = gSoInfoFreeList;
gSoInfoFreeList = gSoInfoFreeList->next;
// Initialize the new element.
memset(si, 0, sizeof(soinfo));
strlcpy(si->name, name, sizeof(si->name));
sonext->next = si;
sonext = si;
TRACE("name %s: allocated soinfo @ %p", name, si);
return si;
}
再回过头来看下soinfo_link_image这个方法，它主要实现了动态链接库中section信息的解析：
- 先解析dynamic section动态节区，进而实现各个Section的定位；
- 获取其他Section的信息；
- 待所有section信息解析完毕后，对HASH,STRTAB,SYMTAB节是否正常解析做校验；
- 若标志位有FLAG_EXE，则表示当前程序执行的是一个可执行文件。到这里可以确定，linker不仅负责加载so，也负责解析加载一个可执行的ELF文件；
- 加载所需要的其他共享库，其中find_library会递归调用这个so_link_image函数，直到某个so库没有DT_NEEDED段；
完成rel节的重定位；
最后，CallConstructors函数会根据动态节区中的信息，获取该共享库所依赖的所有so文件名，并在已加载的动态链接库链表中进行查找、递归调用它们的初始化函数。当运行所需的依赖库都初始化完成后，再执行init_func、init_array方法初始化该动态库。。void soinfo::CallConstructors() { if (constructors_called) { return; } // We set constructors_called before actually calling the constructors, otherwise it doesn't // protect against recursive constructor calls. One simple example of constructor recursion // is the libc debug malloc, which is implemented in libc_malloc_debug_leak.so: // 1. The program depends on libc, so libc's constructor is called here. // 2. The libc constructor calls dlopen() to load libc_malloc_debug_leak.so. // 3. dlopen() calls the constructors on the newly created // soinfo for libc_malloc_debug_leak.so. // 4. The debug .so depends on libc, so CallConstructors is // called again with the libc soinfo. If it doesn't trigger the early- // out above, the libc constructor will be called again (recursively!). constructors_called = true; if ((flags & FLAG_EXE) == 0 && preinit_array != NULL) { // The GNU dynamic linker silently ignores these, but we warn the developer. PRINT("\"%s\": ignoring %d-entry DT_PREINIT_ARRAY in shared library!", name, preinit_array_count); } if (dynamic != NULL) { for (Elf32_Dyn* d = dynamic; d->d_tag != DT_NULL; ++d) { if (d->d_tag == DT_NEEDED) { const char* library_name = strtab + d->d_un.d_val; TRACE("\"%s\": calling constructors in DT_NEEDED \"%s\"", name, library_name); find_loaded_library(library_name)->CallConstructors(); } } } TRACE("\"%s\": calling constructors", name); // DT_INIT should be called before DT_INIT_ARRAY if both are present. CallFunction("DT_INIT", init_func); CallArray("DT_INIT_ARRAY", init_array, init_array_count, false); }
loadLibrary之加载调用
Java层通过System.load或System.loadLibrary来加载一个so文件，它的定义在Android源码中的路径为/libcore/luni/src/main/java/java/lang/System.java，执行流程如下：
接下来，让我们具体看下System.loadLibrary这个方法的实现。可以发现它实际是先通过VMStack.getCallingClassLoader()获取到ClassLoader，然后调用运行时的loadLibrary。
/**
* Loads and links the library with the specified name. The mapping of the
* specified library name to the full path for loading the library is
* implementation-dependent.
*
* @param libName
* the name of the library to load.
* @throws UnsatisfiedLinkError
* if the library can not be loaded.
*/
public void loadLibrary(String libName) {
loadLibrary(libName, VMStack.getCallingClassLoader());
}
/*
* Searches for a library, then loads and links it without security checks.
*/
void loadLibrary(String libraryName, ClassLoader loader) {
if (loader != null) {
String filename = loader.findLibrary(libraryName);
if (filename == null) {
throw new UnsatisfiedLinkError("Couldn't load " + libraryName +
" from loader " + loader +
": findLibrary returned null");
}
String error = doLoad(filename, loader);
if (error != null) {
throw new UnsatisfiedLinkError(error);
}
return;
}
String filename = System.mapLibraryName(libraryName);
List<String> candidates = new ArrayList<String>();
String lastError = null;
for (String directory : mLibPaths) {
String candidate = directory + filename;
candidates.add(candidate);
if (IoUtils.canOpenReadOnly(candidate)) {
String error = doLoad(candidate, loader);
if (error == null) {
return; // We successfully loaded the library. Job done.
}
lastError = error;
}
}
if (lastError != null) {
throw new UnsatisfiedLinkError(lastError);
}
throw new UnsatisfiedLinkError("Library " + libraryName + " not found; tried " + candidates);
}
以上代码块的主要功能为：
- 若ClassLoader非空，则利用ClassLoader的findLibrary方法来获取library的path；
- 若ClassLoader为空，则根据传递进来的libraryName，获取到library file的name（比如传递“test”进来，经过System.mapLibraryName方法的调用，返回的会是“libtest.so”）。然后再在一个path list（即下面代码截图中的mLibPaths）中查找到这个library file，并最终确定library 的path；
- 调用nativeLoad这个jni方法来load library。
注意：上传附件及图片大小不得大于30M。
⚠️ 版权声明：
本博客所有内容（含教程、源码、工具）仅供个人技术学习与研究交流使用，严禁商用、倒卖、二次分发及非法用途。
未经作者书面授权，任何组织或个人不得转载、复制或用于其他平台，违者将追究相关责任。
参与讨论
登录后可查看全部回复并参与讨论。
