gecko-dev/security/sandbox/linux/Sandbox.cpp

689 строки
22 KiB
C++

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this file,
* You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "Sandbox.h"
#include "LinuxCapabilities.h"
#include "LinuxSched.h"
#include "SandboxBrokerClient.h"
#include "SandboxChroot.h"
#include "SandboxFilter.h"
#include "SandboxInternal.h"
#include "SandboxLogging.h"
#include "SandboxUtil.h"
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <linux/futex.h>
#include <pthread.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/prctl.h>
#include <sys/ptrace.h>
#include <sys/syscall.h>
#include <sys/time.h>
#include <unistd.h>
#include "mozilla/Atomics.h"
#include "mozilla/Maybe.h"
#include "mozilla/SandboxInfo.h"
#include "mozilla/UniquePtr.h"
#include "mozilla/unused.h"
#include "sandbox/linux/bpf_dsl/dump_bpf.h"
#include "sandbox/linux/bpf_dsl/policy.h"
#include "sandbox/linux/bpf_dsl/policy_compiler.h"
#include "sandbox/linux/seccomp-bpf/linux_seccomp.h"
#include "sandbox/linux/seccomp-bpf/trap.h"
#if defined(ANDROID)
#include "sandbox/linux/services/android_ucontext.h"
#endif
#include "sandbox/linux/services/linux_syscalls.h"
#ifdef MOZ_ASAN
// Copy libsanitizer declarations to avoid depending on ASAN headers.
// See also bug 1081242 comment #4.
extern "C" {
namespace __sanitizer {
// Win64 uses long long, but this is Linux.
typedef signed long sptr;
} // namespace __sanitizer
typedef struct {
int coverage_sandboxed;
__sanitizer::sptr coverage_fd;
unsigned int coverage_max_block_size;
} __sanitizer_sandbox_arguments;
MOZ_IMPORT_API void
__sanitizer_sandbox_on_notify(__sanitizer_sandbox_arguments *args);
} // extern "C"
#endif // MOZ_ASAN
namespace mozilla {
#ifdef ANDROID
SandboxCrashFunc gSandboxCrashFunc;
#endif
#ifdef MOZ_GMP_SANDBOX
// For media plugins, we can start the sandbox before we dlopen the
// module, so we have to pre-open the file and simulate the sandboxed
// open().
static SandboxOpenedFile gMediaPluginFile;
#endif
static UniquePtr<SandboxChroot> gChrootHelper;
static void (*gChromiumSigSysHandler)(int, siginfo_t*, void*);
// Test whether a ucontext, interpreted as the state after a syscall,
// indicates the given error. See also sandbox::Syscall::PutValueInUcontext.
static bool
ContextIsError(const ucontext_t *aContext, int aError)
{
// Avoid integer promotion warnings. (The unary addition makes
// the decltype not evaluate to a reference type.)
typedef decltype(+SECCOMP_RESULT(aContext)) reg_t;
#ifdef __mips__
return SECCOMP_PARM4(aContext) != 0
&& SECCOMP_RESULT(aContext) == static_cast<reg_t>(aError);
#else
return SECCOMP_RESULT(aContext) == static_cast<reg_t>(-aError);
#endif
}
/**
* This is the SIGSYS handler function. It delegates to the Chromium
* TrapRegistry handler (see InstallSigSysHandler, below) and, if the
* trap handler installed by the policy would fail with ENOSYS,
* crashes the process. This allows unintentional policy failures to
* be reported as crash dumps and fixed. It also logs information
* about the failed system call.
*
* Note that this could be invoked in parallel on multiple threads and
* that it could be in async signal context (e.g., intercepting an
* open() called from an async signal handler).
*/
static void
SigSysHandler(int nr, siginfo_t *info, void *void_context)
{
ucontext_t *ctx = static_cast<ucontext_t*>(void_context);
// This shouldn't ever be null, but the Chromium handler checks for
// that and refrains from crashing, so let's not crash release builds:
MOZ_DIAGNOSTIC_ASSERT(ctx);
if (!ctx) {
return;
}
// Save a copy of the context before invoking the trap handler,
// which will overwrite one or more registers with the return value.
ucontext_t savedCtx = *ctx;
gChromiumSigSysHandler(nr, info, ctx);
if (!ContextIsError(ctx, ENOSYS)) {
return;
}
pid_t pid = getpid();
unsigned long syscall_nr = SECCOMP_SYSCALL(&savedCtx);
unsigned long args[6];
args[0] = SECCOMP_PARM1(&savedCtx);
args[1] = SECCOMP_PARM2(&savedCtx);
args[2] = SECCOMP_PARM3(&savedCtx);
args[3] = SECCOMP_PARM4(&savedCtx);
args[4] = SECCOMP_PARM5(&savedCtx);
args[5] = SECCOMP_PARM6(&savedCtx);
// TODO, someday when this is enabled on MIPS: include the two extra
// args in the error message.
SANDBOX_LOG_ERROR("seccomp sandbox violation: pid %d, syscall %d,"
" args %d %d %d %d %d %d. Killing process.",
pid, syscall_nr,
args[0], args[1], args[2], args[3], args[4], args[5]);
// Bug 1017393: record syscall number somewhere useful.
info->si_addr = reinterpret_cast<void*>(syscall_nr);
gSandboxCrashFunc(nr, info, &savedCtx);
_exit(127);
}
/**
* This function installs the SIGSYS handler. This is slightly
* complicated because we want to use Chromium's handler to dispatch
* to specific trap handlers defined in the policy, but we also need
* the full original signal context to give to Breakpad for crash
* dumps. So we install Chromium's handler first, then retrieve its
* address so our replacement can delegate to it.
*/
static void
InstallSigSysHandler(void)
{
struct sigaction act;
// Ensure that the Chromium handler is installed.
Unused << sandbox::Trap::Registry();
// If the signal handling state isn't as expected, crash now instead
// of crashing later (and more confusingly) when SIGSYS happens.
if (sigaction(SIGSYS, nullptr, &act) != 0) {
MOZ_CRASH("Couldn't read old SIGSYS disposition");
}
if ((act.sa_flags & SA_SIGINFO) != SA_SIGINFO) {
MOZ_CRASH("SIGSYS not already set to a siginfo handler?");
}
MOZ_RELEASE_ASSERT(act.sa_sigaction);
gChromiumSigSysHandler = act.sa_sigaction;
act.sa_sigaction = SigSysHandler;
// Currently, SA_NODEFER should already be set by the Chromium code,
// but it's harmless to ensure that it's set:
MOZ_ASSERT(act.sa_flags & SA_NODEFER);
act.sa_flags |= SA_NODEFER;
if (sigaction(SIGSYS, &act, nullptr) < 0) {
MOZ_CRASH("Couldn't change SIGSYS disposition");
}
}
/**
* This function installs the syscall filter, a.k.a. seccomp. The
* aUseTSync flag indicates whether this should apply to all threads
* in the process -- which will fail if the kernel doesn't support
* that -- or only the current thread.
*
* SECCOMP_MODE_FILTER is the "bpf" mode of seccomp which allows
* to pass a bpf program (in our case, it contains a syscall
* whitelist).
*
* PR_SET_NO_NEW_PRIVS ensures that it is impossible to grant more
* syscalls to the process beyond this point (even after fork()), and
* prevents gaining capabilities (e.g., by exec'ing a setuid root
* program). The kernel won't allow seccomp-bpf without doing this,
* because otherwise it could be used for privilege escalation attacks.
*
* Returns false (and sets errno) on failure.
*
* @see SandboxInfo
* @see BroadcastSetThreadSandbox
*/
static bool MOZ_MUST_USE
InstallSyscallFilter(const sock_fprog *aProg, bool aUseTSync)
{
if (prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0)) {
SANDBOX_LOG_ERROR("prctl(PR_SET_NO_NEW_PRIVS) failed: %s", strerror(errno));
MOZ_CRASH("prctl(PR_SET_NO_NEW_PRIVS)");
}
if (aUseTSync) {
if (syscall(__NR_seccomp, SECCOMP_SET_MODE_FILTER,
SECCOMP_FILTER_FLAG_TSYNC, aProg) != 0) {
SANDBOX_LOG_ERROR("thread-synchronized seccomp failed: %s",
strerror(errno));
return false;
}
} else {
if (prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, (unsigned long)aProg, 0, 0)) {
SANDBOX_LOG_ERROR("prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER) failed: %s",
strerror(errno));
return false;
}
}
return true;
}
// Use signals for permissions that need to be set per-thread.
// The communication channel from the signal handler back to the main thread.
static mozilla::Atomic<int> gSetSandboxDone;
// Pass the filter itself through a global.
static const sock_fprog* gSetSandboxFilter;
// We have to dynamically allocate the signal number; see bug 1038900.
// This function returns the first realtime signal currently set to
// default handling (i.e., not in use), or 0 if none could be found.
//
// WARNING: if this function or anything similar to it (including in
// external libraries) is used on multiple threads concurrently, there
// will be a race condition.
static int
FindFreeSignalNumber()
{
for (int signum = SIGRTMIN; signum <= SIGRTMAX; ++signum) {
struct sigaction sa;
if (sigaction(signum, nullptr, &sa) == 0 &&
(sa.sa_flags & SA_SIGINFO) == 0 &&
sa.sa_handler == SIG_DFL) {
return signum;
}
}
return 0;
}
// Returns true if sandboxing was enabled, or false if sandboxing
// already was enabled. Crashes if sandboxing could not be enabled.
static bool
SetThreadSandbox()
{
if (prctl(PR_GET_SECCOMP, 0, 0, 0, 0) == 0) {
if (!InstallSyscallFilter(gSetSandboxFilter, false)) {
MOZ_CRASH("prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER)");
}
return true;
}
return false;
}
static void
SetThreadSandboxHandler(int signum)
{
// The non-zero number sent back to the main thread indicates
// whether action was taken.
if (SetThreadSandbox()) {
gSetSandboxDone = 2;
} else {
gSetSandboxDone = 1;
}
// Wake up the main thread. See the FUTEX_WAIT call, below, for an
// explanation.
syscall(__NR_futex, reinterpret_cast<int*>(&gSetSandboxDone),
FUTEX_WAKE, 1);
}
static void
EnterChroot()
{
if (gChrootHelper) {
gChrootHelper->Invoke();
gChrootHelper = nullptr;
}
}
static void
BroadcastSetThreadSandbox(const sock_fprog* aFilter)
{
int signum;
pid_t pid, tid, myTid;
DIR *taskdp;
struct dirent *de;
// This function does not own *aFilter, so this global needs to
// always be zeroed before returning.
gSetSandboxFilter = aFilter;
static_assert(sizeof(mozilla::Atomic<int>) == sizeof(int),
"mozilla::Atomic<int> isn't represented by an int");
pid = getpid();
myTid = syscall(__NR_gettid);
taskdp = opendir("/proc/self/task");
if (taskdp == nullptr) {
SANDBOX_LOG_ERROR("opendir /proc/self/task: %s\n", strerror(errno));
MOZ_CRASH();
}
EnterChroot();
signum = FindFreeSignalNumber();
if (signum == 0) {
SANDBOX_LOG_ERROR("No available signal numbers!");
MOZ_CRASH();
}
void (*oldHandler)(int);
oldHandler = signal(signum, SetThreadSandboxHandler);
if (oldHandler != SIG_DFL) {
// See the comment on FindFreeSignalNumber about race conditions.
SANDBOX_LOG_ERROR("signal %d in use by handler %p!\n", signum, oldHandler);
MOZ_CRASH();
}
// In case this races with a not-yet-deprivileged thread cloning
// itself, repeat iterating over all threads until we find none
// that are still privileged.
bool sandboxProgress;
do {
sandboxProgress = false;
// For each thread...
while ((de = readdir(taskdp))) {
char *endptr;
tid = strtol(de->d_name, &endptr, 10);
if (*endptr != '\0' || tid <= 0) {
// Not a task ID.
continue;
}
if (tid == myTid) {
// Drop this thread's privileges last, below, so we can
// continue to signal other threads.
continue;
}
// Reset the futex cell and signal.
gSetSandboxDone = 0;
if (syscall(__NR_tgkill, pid, tid, signum) != 0) {
if (errno == ESRCH) {
SANDBOX_LOG_ERROR("Thread %d unexpectedly exited.", tid);
// Rescan threads, in case it forked before exiting.
sandboxProgress = true;
continue;
}
SANDBOX_LOG_ERROR("tgkill(%d,%d): %s\n", pid, tid, strerror(errno));
MOZ_CRASH();
}
// It's unlikely, but if the thread somehow manages to exit
// after receiving the signal but before entering the signal
// handler, we need to avoid blocking forever.
//
// Using futex directly lets the signal handler send the wakeup
// from an async signal handler (pthread mutex/condvar calls
// aren't allowed), and to use a relative timeout that isn't
// affected by changes to the system clock (not possible with
// POSIX semaphores).
//
// If a thread doesn't respond within a reasonable amount of
// time, but still exists, we crash -- the alternative is either
// blocking forever or silently losing security, and it
// shouldn't actually happen.
static const int crashDelay = 10; // seconds
struct timespec timeLimit;
clock_gettime(CLOCK_MONOTONIC, &timeLimit);
timeLimit.tv_sec += crashDelay;
while (true) {
static const struct timespec futexTimeout = { 0, 10*1000*1000 }; // 10ms
// Atomically: if gSetSandboxDone == 0, then sleep.
if (syscall(__NR_futex, reinterpret_cast<int*>(&gSetSandboxDone),
FUTEX_WAIT, 0, &futexTimeout) != 0) {
if (errno != EWOULDBLOCK && errno != ETIMEDOUT && errno != EINTR) {
SANDBOX_LOG_ERROR("FUTEX_WAIT: %s\n", strerror(errno));
MOZ_CRASH();
}
}
// Did the handler finish?
if (gSetSandboxDone > 0) {
if (gSetSandboxDone == 2) {
sandboxProgress = true;
}
break;
}
// Has the thread ceased to exist?
if (syscall(__NR_tgkill, pid, tid, 0) != 0) {
if (errno == ESRCH) {
SANDBOX_LOG_ERROR("Thread %d unexpectedly exited.", tid);
}
// Rescan threads, in case it forked before exiting.
// Also, if it somehow failed in a way that wasn't ESRCH,
// and still exists, that will be handled on the next pass.
sandboxProgress = true;
break;
}
struct timespec now;
clock_gettime(CLOCK_MONOTONIC, &now);
if (now.tv_sec > timeLimit.tv_sec ||
(now.tv_sec == timeLimit.tv_sec &&
now.tv_nsec > timeLimit.tv_nsec)) {
SANDBOX_LOG_ERROR("Thread %d unresponsive for %d seconds."
" Killing process.",
tid, crashDelay);
MOZ_CRASH();
}
}
}
rewinddir(taskdp);
} while (sandboxProgress);
oldHandler = signal(signum, SIG_DFL);
if (oldHandler != SetThreadSandboxHandler) {
// See the comment on FindFreeSignalNumber about race conditions.
SANDBOX_LOG_ERROR("handler for signal %d was changed to %p!",
signum, oldHandler);
MOZ_CRASH();
}
Unused << closedir(taskdp);
// And now, deprivilege the main thread:
SetThreadSandbox();
gSetSandboxFilter = nullptr;
}
static void
ApplySandboxWithTSync(sock_fprog* aFilter)
{
EnterChroot();
// At this point we're committed to using tsync, because the signal
// broadcast workaround needs to access procfs. (Unless chroot
// isn't used... but this failure shouldn't happen in the first
// place, so let's not make extra special cases for it.)
if (!InstallSyscallFilter(aFilter, true)) {
MOZ_CRASH("seccomp+tsync failed, but kernel supports tsync");
}
}
// Common code for sandbox startup.
static void
SetCurrentProcessSandbox(UniquePtr<sandbox::bpf_dsl::Policy> aPolicy)
{
MOZ_ASSERT(gSandboxCrashFunc);
// Note: PolicyCompiler borrows the policy and registry for its
// lifetime, but does not take ownership of them.
sandbox::bpf_dsl::PolicyCompiler compiler(aPolicy.get(),
sandbox::Trap::Registry());
auto program = compiler.Compile();
if (SandboxInfo::Get().Test(SandboxInfo::kVerbose)) {
sandbox::bpf_dsl::DumpBPF::PrintProgram(*program);
}
InstallSigSysHandler();
#ifdef MOZ_ASAN
__sanitizer_sandbox_arguments asanArgs;
asanArgs.coverage_sandboxed = 1;
asanArgs.coverage_fd = -1;
asanArgs.coverage_max_block_size = 0;
__sanitizer_sandbox_on_notify(&asanArgs);
#endif
// The syscall takes a C-style array, so copy the vector into one.
size_t programLen = program->size();
UniquePtr<sock_filter[]> flatProgram(new sock_filter[programLen]);
for (auto i = program->begin(); i != program->end(); ++i) {
flatProgram[i - program->begin()] = *i;
}
sock_fprog fprog;
fprog.filter = flatProgram.get();
fprog.len = static_cast<unsigned short>(programLen);
MOZ_RELEASE_ASSERT(static_cast<size_t>(fprog.len) == programLen);
const SandboxInfo info = SandboxInfo::Get();
if (info.Test(SandboxInfo::kHasSeccompTSync)) {
if (info.Test(SandboxInfo::kVerbose)) {
SANDBOX_LOG_ERROR("using seccomp tsync");
}
ApplySandboxWithTSync(&fprog);
} else {
if (info.Test(SandboxInfo::kVerbose)) {
SANDBOX_LOG_ERROR("no tsync support; using signal broadcast");
}
BroadcastSetThreadSandbox(&fprog);
}
MOZ_RELEASE_ASSERT(!gChrootHelper, "forgot to chroot");
}
void
SandboxEarlyInit(GeckoProcessType aType, bool aIsNuwa)
{
// Bug 1168555: Nuwa isn't reliably single-threaded at this point;
// it starts an IPC I/O thread and then shuts it down before calling
// the plugin-container entry point, but that thread may not have
// finished exiting. If/when any type of sandboxing is used for the
// Nuwa process (e.g., unsharing the network namespace there instead
// of for each content process, to save memory), this will need to be
// changed by moving the SandboxEarlyInit call to an earlier point.
if (aIsNuwa) {
return;
}
const SandboxInfo info = SandboxInfo::Get();
if (info.Test(SandboxInfo::kUnexpectedThreads)) {
return;
}
MOZ_RELEASE_ASSERT(IsSingleThreaded());
// Which kinds of resource isolation (of those that need to be set
// up at this point) can be used by this process?
bool canChroot = false;
bool canUnshareNet = false;
bool canUnshareIPC = false;
switch (aType) {
case GeckoProcessType_Default:
MOZ_ASSERT(false, "SandboxEarlyInit in parent process");
return;
#ifdef MOZ_GMP_SANDBOX
case GeckoProcessType_GMPlugin:
if (!info.Test(SandboxInfo::kEnabledForMedia)) {
break;
}
canUnshareNet = true;
canUnshareIPC = true;
// Need seccomp-bpf to intercept open().
canChroot = info.Test(SandboxInfo::kHasSeccompBPF);
break;
#endif
// In the future, content processes will be able to use some of
// these.
default:
// Other cases intentionally left blank.
break;
}
// If there's nothing to do, then we're done.
if (!canChroot && !canUnshareNet && !canUnshareIPC) {
return;
}
{
LinuxCapabilities existingCaps;
if (existingCaps.GetCurrent() && existingCaps.AnyEffective()) {
SANDBOX_LOG_ERROR("PLEASE DO NOT RUN THIS AS ROOT. Strange things may"
" happen when capabilities are dropped.");
}
}
// If capabilities can't be gained, then nothing can be done.
if (!info.Test(SandboxInfo::kHasUserNamespaces)) {
// Drop any existing capabilities; unsharing the user namespace
// would implicitly drop them, so if we're running in a broken
// configuration where that would matter (e.g., running as root
// from a non-root-owned mode-0700 directory) this means it will
// break the same way on all kernels and be easier to troubleshoot.
LinuxCapabilities().SetCurrent();
return;
}
// The failure cases for the various unshares, and setting up the
// chroot helper, don't strictly need to be fatal -- but they also
// shouldn't fail on any reasonable system, so let's take the small
// risk of breakage over the small risk of quietly providing less
// security than we expect. (Unlike in SandboxInfo, this is in the
// child process, so crashing here isn't as severe a response to the
// unexpected.)
if (!UnshareUserNamespace()) {
SANDBOX_LOG_ERROR("unshare(CLONE_NEWUSER): %s", strerror(errno));
// If CanCreateUserNamespace (SandboxInfo.cpp) returns true, then
// the unshare shouldn't have failed.
MOZ_CRASH("unshare(CLONE_NEWUSER)");
}
// No early returns after this point! We need to drop the
// capabilities that were gained by unsharing the user namesapce.
if (canUnshareIPC && syscall(__NR_unshare, CLONE_NEWIPC) != 0) {
SANDBOX_LOG_ERROR("unshare(CLONE_NEWIPC): %s", strerror(errno));
MOZ_CRASH("unshare(CLONE_NEWIPC)");
}
if (canUnshareNet && syscall(__NR_unshare, CLONE_NEWNET) != 0) {
SANDBOX_LOG_ERROR("unshare(CLONE_NEWNET): %s", strerror(errno));
MOZ_CRASH("unshare(CLONE_NEWNET)");
}
if (canChroot) {
gChrootHelper = MakeUnique<SandboxChroot>();
if (!gChrootHelper->Prepare()) {
SANDBOX_LOG_ERROR("failed to set up chroot helper");
MOZ_CRASH("SandboxChroot::Prepare");
}
}
if (!LinuxCapabilities().SetCurrent()) {
SANDBOX_LOG_ERROR("dropping capabilities: %s", strerror(errno));
MOZ_CRASH("can't drop capabilities");
}
}
#ifdef MOZ_CONTENT_SANDBOX
/**
* Starts the seccomp sandbox for a content process. Should be called
* only once, and before any potentially harmful content is loaded.
*
* Will normally make the process exit on failure.
*/
void
SetContentProcessSandbox(int aBrokerFd)
{
if (!SandboxInfo::Get().Test(SandboxInfo::kEnabledForContent)) {
if (aBrokerFd >= 0) {
close(aBrokerFd);
}
return;
}
// This needs to live until the process exits.
static Maybe<SandboxBrokerClient> sBroker;
if (aBrokerFd >= 0) {
sBroker.emplace(aBrokerFd);
}
SetCurrentProcessSandbox(GetContentSandboxPolicy(sBroker.ptrOr(nullptr)));
}
#endif // MOZ_CONTENT_SANDBOX
#ifdef MOZ_GMP_SANDBOX
/**
* Starts the seccomp sandbox for a media plugin process. Should be
* called only once, and before any potentially harmful content is
* loaded -- including the plugin itself, if it's considered untrusted.
*
* The file indicated by aFilePath, if non-null, can be open()ed
* read-only, once, after the sandbox starts; it should be the .so
* file implementing the not-yet-loaded plugin.
*
* Will normally make the process exit on failure.
*/
void
SetMediaPluginSandbox(const char *aFilePath)
{
if (!SandboxInfo::Get().Test(SandboxInfo::kEnabledForMedia)) {
return;
}
MOZ_ASSERT(!gMediaPluginFile.mPath);
if (aFilePath) {
gMediaPluginFile.mPath = strdup(aFilePath);
gMediaPluginFile.mFd = open(aFilePath, O_RDONLY | O_CLOEXEC);
if (gMediaPluginFile.mFd == -1) {
SANDBOX_LOG_ERROR("failed to open plugin file %s: %s",
aFilePath, strerror(errno));
MOZ_CRASH();
}
} else {
gMediaPluginFile.mFd = -1;
}
// Finally, start the sandbox.
SetCurrentProcessSandbox(GetMediaSandboxPolicy(&gMediaPluginFile));
}
#endif // MOZ_GMP_SANDBOX
} // namespace mozilla