platform_bionic/libc/bionic/libc_init_common.cpp

370 lines
11 KiB
C++
Raw Normal View History

/*
* Copyright (C) 2008 The Android Open Source Project
* 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.
*
* 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.
*/
#include "libc_init_common.h"
#include <elf.h>
#include <errno.h>
#include <fcntl.h>
#include <stddef.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/auxv.h>
#include <sys/personality.h>
#include <sys/time.h>
#include <unistd.h>
#include <async_safe/log.h>
#include "private/WriteProtected.h"
#include "private/bionic_defs.h"
#include "private/bionic_globals.h"
#include "private/bionic_tls.h"
#include "private/thread_private.h"
#include "pthread_internal.h"
extern "C" int __system_properties_init(void);
__LIBC_HIDDEN__ WriteProtected<libc_globals> __libc_globals;
// Not public, but well-known in the BSDs.
const char* __progname;
void __libc_init_globals() {
// Initialize libc globals that are needed in both the linker and in libc.
// In dynamic binaries, this is run at least twice for different copies of the
// globals, once for the linker's copy and once for the one in libc.so.
__libc_globals.initialize();
__libc_globals.mutate([](libc_globals* globals) {
__libc_init_vdso(globals);
__libc_init_setjmp_cookie(globals);
});
}
#if !defined(__LP64__)
static void __check_max_thread_id() {
if (gettid() > 65535) {
async_safe_fatal("Limited by the size of pthread_mutex_t, 32 bit bionic libc only accepts "
"pid <= 65535, but current pid is %d", gettid());
}
}
#endif
static void arc4random_fork_handler() {
_rs_forked = 1;
_thread_arc4_lock();
}
__BIONIC_WEAK_FOR_NATIVE_BRIDGE
void __libc_add_main_thread() {
// Get the main thread from TLS and add it to the thread list.
pthread_internal_t* main_thread = __get_thread();
__pthread_internal_add(main_thread);
}
void __libc_init_common() {
// Initialize various globals.
environ = __libc_shared_globals()->init_environ;
errno = 0;
__progname = __libc_shared_globals()->init_progname ?: "<unknown>";
#if !defined(__LP64__)
__check_max_thread_id();
#endif
__libc_add_main_thread();
// Register atfork handlers to take and release the arc4random lock.
pthread_atfork(arc4random_fork_handler, _thread_arc4_unlock, _thread_arc4_unlock);
__system_properties_init(); // Requires 'environ'.
__libc_init_fdsan(); // Requires system properties (for debug.fdsan).
}
__noreturn static void __early_abort(int line) {
// We can't write to stdout or stderr because we're aborting before we've checked that
// it's safe for us to use those file descriptors. We probably can't strace either, so
// we rely on the fact that if we dereference a low address, either debuggerd or the
// kernel's crash dump will show the fault address.
*reinterpret_cast<int*>(line) = 0;
_exit(EXIT_FAILURE);
}
// Force any of the closed stdin, stdout and stderr to be associated with /dev/null.
static void __nullify_closed_stdio() {
int dev_null = TEMP_FAILURE_RETRY(open("/dev/null", O_RDWR));
if (dev_null == -1) {
// init won't have /dev/null available, but SELinux provides an equivalent.
dev_null = TEMP_FAILURE_RETRY(open("/sys/fs/selinux/null", O_RDWR));
}
if (dev_null == -1) {
__early_abort(__LINE__);
}
// If any of the stdio file descriptors is valid and not associated
// with /dev/null, dup /dev/null to it.
for (int i = 0; i < 3; i++) {
// If it is /dev/null already, we are done.
if (i == dev_null) {
continue;
}
// Is this fd already open?
int status = TEMP_FAILURE_RETRY(fcntl(i, F_GETFL));
if (status != -1) {
continue;
}
// The only error we allow is that the file descriptor does not
// exist, in which case we dup /dev/null to it.
if (errno == EBADF) {
// Try dupping /dev/null to this stdio file descriptor and
// repeat if there is a signal. Note that any errors in closing
// the stdio descriptor are lost.
status = TEMP_FAILURE_RETRY(dup2(dev_null, i));
if (status == -1) {
__early_abort(__LINE__);
}
} else {
__early_abort(__LINE__);
}
}
// If /dev/null is not one of the stdio file descriptors, close it.
if (dev_null > 2) {
if (close(dev_null) == -1) {
__early_abort(__LINE__);
}
}
}
// Check if the environment variable definition at 'envstr'
// starts with '<name>=', and if so return the address of the
// first character after the equal sign. Otherwise return null.
static const char* env_match(const char* envstr, const char* name) {
size_t i = 0;
while (envstr[i] == name[i] && name[i] != '\0') {
++i;
}
if (name[i] == '\0' && envstr[i] == '=') {
return envstr + i + 1;
}
return nullptr;
}
static bool __is_valid_environment_variable(const char* name) {
// According to the kernel source, by default the kernel uses 32*PAGE_SIZE
// as the maximum size for an environment variable definition.
const int MAX_ENV_LEN = 32*4096;
if (name == nullptr) {
return false;
}
// Parse the string, looking for the first '=' there, and its size.
int pos = 0;
int first_equal_pos = -1;
while (pos < MAX_ENV_LEN) {
if (name[pos] == '\0') {
break;
}
if (name[pos] == '=' && first_equal_pos < 0) {
first_equal_pos = pos;
}
pos++;
}
// Check that it's smaller than MAX_ENV_LEN (to detect non-zero terminated strings).
if (pos >= MAX_ENV_LEN) {
return false;
}
// Check that it contains at least one equal sign that is not the first character
if (first_equal_pos < 1) {
return false;
}
return true;
}
static bool __is_unsafe_environment_variable(const char* name) {
// None of these should be allowed when the AT_SECURE auxv
// flag is set. This flag is set to inform userspace that a
// security transition has occurred, for example, as a result
// of executing a setuid program or the result of an SELinux
// security transition.
static constexpr const char* UNSAFE_VARIABLE_NAMES[] = {
"ANDROID_DNS_MODE",
"GCONV_PATH",
"GETCONF_DIR",
"HOSTALIASES",
"JE_MALLOC_CONF",
"LD_AOUT_LIBRARY_PATH",
"LD_AOUT_PRELOAD",
"LD_AUDIT",
Don't honor LD_CONFIG_FILE across security transitions For security reasons, when a binary is executed which causes a security transition (eg, a setuid binary, setgid binary, filesystem capabilities, or SELinux domain transition), the AT_SECURE flag is set. This causes certain blacklisted environment variables to be stripped before the process is executed. The list of blacklisted environment variables is stored in UNSAFE_VARIABLE_NAMES. Generally speaking, most environment variables used internally by libc show up in this list. Commit 02586a2a34e6acfccf359b94db840f422b6c0231 ("linker: the global group is added to all built-in namespaces", Aug 2017) added support for the environment variable LD_CONFIG_FILE. This debug build only feature allows the caller to specify the path to the loader configuration file. Like other linker environment variables, setting this variable allows the calling process to control executed code of the called process, which has security implications (on debuggable builds only). Add LD_CONFIG_FILE to UNSAFE_VARIABLE_NAMES. This has the effect of stripping, on all build types, the LD_CONFIG_FILE environment variable. This has three advantages: 1) Prevents security bugs should LD_CONFIG_FILE ever be inadvertantly exposed on a production build. 2) Makes the behavior of userdebug and user builds more similar, helping prevent build-type dependent bugs where someone may come to rely on this debug-only feature. 3) Protect droidfood users against malicious applications which can trigger a security transition, eg, the execution of crash_dump or the renderscript compiler. Alternative considered but rejected: If we treated LD_CONFIG_FILE like LD_PRELOAD, we could expose this on all build types, and remove the build-type dependent behavior. But this is contrary to enh's Aug 02 2017 guidance at https://android-review.googlesource.com/c/platform/bionic/+/449956 i'm still uncomfortable about LD_CONFIG_FILE because i'd like to be reducing the number of environment variables that affect the linker in P rather than increasing them. Test: atest CtsBionicTestCases Test: atest linker-unit-tests Change-Id: I82d286546ee079b5cde04428dc89941c253c2d20
2019-04-26 20:25:02 +02:00
"LD_CONFIG_FILE",
"LD_DEBUG",
"LD_DEBUG_OUTPUT",
"LD_DYNAMIC_WEAK",
"LD_LIBRARY_PATH",
"LD_ORIGIN_PATH",
"LD_PRELOAD",
"LD_PROFILE",
"LD_SHOW_AUXV",
"LD_USE_LOAD_BIAS",
"LIBC_DEBUG_MALLOC_OPTIONS",
"LIBC_HOOKS_ENABLE",
"LOCALDOMAIN",
"LOCPATH",
"MALLOC_CHECK_",
"MALLOC_CONF",
"MALLOC_TRACE",
"NIS_PATH",
"NLSPATH",
"RESOLV_HOST_CONF",
"RES_OPTIONS",
"TMPDIR",
"TZDIR",
};
for (const auto& unsafe_variable_name : UNSAFE_VARIABLE_NAMES) {
if (env_match(name, unsafe_variable_name) != nullptr) {
return true;
}
}
return false;
}
static void __sanitize_environment_variables(char** env) {
char** src = env;
char** dst = env;
for (; src[0] != nullptr; ++src) {
if (!__is_valid_environment_variable(src[0])) {
continue;
}
// Remove various unsafe environment variables if we're loading a setuid program.
if (__is_unsafe_environment_variable(src[0])) {
continue;
}
dst[0] = src[0];
++dst;
}
dst[0] = nullptr;
}
static void __initialize_personality() {
#if !defined(__LP64__)
int old_value = personality(0xffffffff);
if (old_value == -1) {
async_safe_fatal("error getting old personality value: %s", strerror(errno));
}
if (personality((static_cast<unsigned int>(old_value) & ~PER_MASK) | PER_LINUX32) == -1) {
async_safe_fatal("error setting PER_LINUX32 personality: %s", strerror(errno));
}
#endif
}
void __libc_init_AT_SECURE(char** env) {
// Check that the kernel provided a value for AT_SECURE.
errno = 0;
unsigned long is_AT_SECURE = getauxval(AT_SECURE);
if (errno != 0) __early_abort(__LINE__);
Ensure STDIN/STDOUT/STDERR always exist File descriptor confusion can result if a process is exec()d and STDIN/STDOUT/STDERR do not exist. In those situations, the first, second, and third files opened by the exec()d application will have FD 0, 1, and 2 respectively. Code which reads / writes to these STD* file descriptors may end up reading / writing to unintended files. To prevent this, guarantee that FDs 0, 1, and 2 always exist. Bionic only currently guarantees this for AT_SECURE programs (eg, a setuid binary, setgid binary, filesystem capabilities, or SELinux domain transition). Extending this to all exec()s adds robustness against this class of bugs. Additionally, it allows a caller to do: close(STDIN_FILENO); close(STDOUT_FILENO); close(STDERR_FILENO); and know that the exec()d process will reopen these file descriptors on its own. This has the potential to simplify other parts of Android, eg https://android-review.googlesource.com/c/platform/system/apex/+/915694 Steps to reproduce: sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID Expected: $ sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID [1] 3154 total 0 dr-x------ 2 shell shell 0 1970-04-17 12:15 . dr-xr-xr-x 9 shell shell 0 1970-04-17 12:15 .. lrwx------ 1 shell shell 64 1970-04-17 12:15 0 -> /dev/null lrwx------ 1 shell shell 64 1970-04-17 12:15 1 -> /dev/null lrwx------ 1 shell shell 64 1970-04-17 12:15 2 -> /dev/null $ [1] + Terminated \sleep 100 <&- >&- 2>&- Actual: $ sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID [1] 16345 total 0 dr-x------ 2 shell shell 0 2019-02-28 20:22 . dr-xr-xr-x 9 shell shell 0 2019-02-28 20:22 .. $ [1] + Terminated \sleep 100 <&- >&- 2>&- Test: manual (see above) Change-Id: I3e05700a1e8ebc7fc9d192211dd9fc030cc40139
2019-02-28 21:57:24 +01:00
// Always ensure that STDIN/STDOUT/STDERR exist. This prevents file
// descriptor confusion bugs where a parent process closes
// STD*, the exec()d process calls open() for an unrelated reason,
// the newly created file descriptor is assigned
// 0<=FD<=2, and unrelated code attempts to read / write to the STD*
// FDs.
// In particular, this can be a security bug for setuid/setgid programs.
// For example:
// https://www.freebsd.org/security/advisories/FreeBSD-SA-02:23.stdio.asc
// However, for robustness reasons, we don't limit these protections to
// just security critical executables.
//
// Init is excluded from these protections unless AT_SECURE is set, as
// /dev/null and/or /sys/fs/selinux/null will not be available at
// early boot.
if ((getpid() != 1) || is_AT_SECURE) {
__nullify_closed_stdio();
Ensure STDIN/STDOUT/STDERR always exist File descriptor confusion can result if a process is exec()d and STDIN/STDOUT/STDERR do not exist. In those situations, the first, second, and third files opened by the exec()d application will have FD 0, 1, and 2 respectively. Code which reads / writes to these STD* file descriptors may end up reading / writing to unintended files. To prevent this, guarantee that FDs 0, 1, and 2 always exist. Bionic only currently guarantees this for AT_SECURE programs (eg, a setuid binary, setgid binary, filesystem capabilities, or SELinux domain transition). Extending this to all exec()s adds robustness against this class of bugs. Additionally, it allows a caller to do: close(STDIN_FILENO); close(STDOUT_FILENO); close(STDERR_FILENO); and know that the exec()d process will reopen these file descriptors on its own. This has the potential to simplify other parts of Android, eg https://android-review.googlesource.com/c/platform/system/apex/+/915694 Steps to reproduce: sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID Expected: $ sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID [1] 3154 total 0 dr-x------ 2 shell shell 0 1970-04-17 12:15 . dr-xr-xr-x 9 shell shell 0 1970-04-17 12:15 .. lrwx------ 1 shell shell 64 1970-04-17 12:15 0 -> /dev/null lrwx------ 1 shell shell 64 1970-04-17 12:15 1 -> /dev/null lrwx------ 1 shell shell 64 1970-04-17 12:15 2 -> /dev/null $ [1] + Terminated \sleep 100 <&- >&- 2>&- Actual: $ sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID [1] 16345 total 0 dr-x------ 2 shell shell 0 2019-02-28 20:22 . dr-xr-xr-x 9 shell shell 0 2019-02-28 20:22 .. $ [1] + Terminated \sleep 100 <&- >&- 2>&- Test: manual (see above) Change-Id: I3e05700a1e8ebc7fc9d192211dd9fc030cc40139
2019-02-28 21:57:24 +01:00
}
Ensure STDIN/STDOUT/STDERR always exist File descriptor confusion can result if a process is exec()d and STDIN/STDOUT/STDERR do not exist. In those situations, the first, second, and third files opened by the exec()d application will have FD 0, 1, and 2 respectively. Code which reads / writes to these STD* file descriptors may end up reading / writing to unintended files. To prevent this, guarantee that FDs 0, 1, and 2 always exist. Bionic only currently guarantees this for AT_SECURE programs (eg, a setuid binary, setgid binary, filesystem capabilities, or SELinux domain transition). Extending this to all exec()s adds robustness against this class of bugs. Additionally, it allows a caller to do: close(STDIN_FILENO); close(STDOUT_FILENO); close(STDERR_FILENO); and know that the exec()d process will reopen these file descriptors on its own. This has the potential to simplify other parts of Android, eg https://android-review.googlesource.com/c/platform/system/apex/+/915694 Steps to reproduce: sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID Expected: $ sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID [1] 3154 total 0 dr-x------ 2 shell shell 0 1970-04-17 12:15 . dr-xr-xr-x 9 shell shell 0 1970-04-17 12:15 .. lrwx------ 1 shell shell 64 1970-04-17 12:15 0 -> /dev/null lrwx------ 1 shell shell 64 1970-04-17 12:15 1 -> /dev/null lrwx------ 1 shell shell 64 1970-04-17 12:15 2 -> /dev/null $ [1] + Terminated \sleep 100 <&- >&- 2>&- Actual: $ sleep 100 <&- >&- 2>&- & BGPID=$! && ls -la /proc/$BGPID/fd && kill $BGPID [1] 16345 total 0 dr-x------ 2 shell shell 0 2019-02-28 20:22 . dr-xr-xr-x 9 shell shell 0 2019-02-28 20:22 .. $ [1] + Terminated \sleep 100 <&- >&- 2>&- Test: manual (see above) Change-Id: I3e05700a1e8ebc7fc9d192211dd9fc030cc40139
2019-02-28 21:57:24 +01:00
if (is_AT_SECURE) {
__sanitize_environment_variables(env);
}
// Now the environment has been sanitized, make it available.
environ = __libc_shared_globals()->init_environ = env;
__initialize_personality();
}
/* This function will be called during normal program termination
* to run the destructors that are listed in the .fini_array section
* of the executable, if any.
*
* 'fini_array' points to a list of function addresses. The first
* entry in the list has value -1, the last one has value 0.
*/
void __libc_fini(void* array) {
typedef void (*Dtor)();
Dtor* fini_array = reinterpret_cast<Dtor*>(array);
const Dtor minus1 = reinterpret_cast<Dtor>(static_cast<uintptr_t>(-1));
// Sanity check - first entry must be -1.
if (array == nullptr || fini_array[0] != minus1) {
return;
}
// Skip over it.
fini_array += 1;
// Count the number of destructors.
int count = 0;
while (fini_array[count] != nullptr) {
++count;
}
// Now call each destructor in reverse order.
while (count > 0) {
Dtor dtor = fini_array[--count];
// Sanity check, any -1 in the list is ignored.
if (dtor == minus1) {
continue;
}
dtor();
}
}