ladybird/Kernel/Thread.h
Idan Horowitz 762e047ec9 Kernel+LibC: Implement sigtimedwait()
This includes a new Thread::Blocker called SignalBlocker which blocks
until a signal of a matching type is pending. The current Blocker
implementation in the Kernel is very complicated, but cleaning it up is
a different yak for a different day.
2021-12-12 08:34:19 +02:00

1480 lines
47 KiB
C++

/*
* Copyright (c) 2018-2021, Andreas Kling <kling@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#pragma once
#include <AK/Concepts.h>
#include <AK/EnumBits.h>
#include <AK/Error.h>
#include <AK/HashMap.h>
#include <AK/IntrusiveList.h>
#include <AK/Optional.h>
#include <AK/OwnPtr.h>
#include <AK/String.h>
#include <AK/TemporaryChange.h>
#include <AK/Time.h>
#include <AK/Variant.h>
#include <AK/Vector.h>
#include <AK/WeakPtr.h>
#include <AK/Weakable.h>
#include <Kernel/Arch/x86/SafeMem.h>
#include <Kernel/Debug.h>
#include <Kernel/FileSystem/InodeIdentifier.h>
#include <Kernel/Forward.h>
#include <Kernel/KString.h>
#include <Kernel/Library/ListedRefCounted.h>
#include <Kernel/Locking/LockLocation.h>
#include <Kernel/Locking/LockMode.h>
#include <Kernel/Locking/LockRank.h>
#include <Kernel/Locking/SpinlockProtected.h>
#include <Kernel/Memory/VirtualRange.h>
#include <Kernel/Scheduler.h>
#include <Kernel/TimerQueue.h>
#include <Kernel/UnixTypes.h>
#include <LibC/fd_set.h>
#include <LibC/signal_numbers.h>
namespace Kernel {
namespace Memory {
extern RecursiveSpinlock s_mm_lock;
}
enum class DispatchSignalResult {
Deferred = 0,
Yield,
Terminate,
Continue
};
struct SignalActionData {
VirtualAddress handler_or_sigaction;
u32 mask { 0 };
int flags { 0 };
};
struct ThreadSpecificData {
ThreadSpecificData* self;
};
#define THREAD_PRIORITY_MIN 1
#define THREAD_PRIORITY_LOW 10
#define THREAD_PRIORITY_NORMAL 30
#define THREAD_PRIORITY_HIGH 50
#define THREAD_PRIORITY_MAX 99
#define THREAD_AFFINITY_DEFAULT 0xffffffff
struct ThreadRegisters {
#if ARCH(I386)
FlatPtr ss;
FlatPtr gs;
FlatPtr fs;
FlatPtr es;
FlatPtr ds;
FlatPtr edi;
FlatPtr esi;
FlatPtr ebp;
FlatPtr esp;
FlatPtr ebx;
FlatPtr edx;
FlatPtr ecx;
FlatPtr eax;
FlatPtr eip;
FlatPtr esp0;
FlatPtr ss0;
#else
FlatPtr rdi;
FlatPtr rsi;
FlatPtr rbp;
FlatPtr rsp;
FlatPtr rbx;
FlatPtr rdx;
FlatPtr rcx;
FlatPtr rax;
FlatPtr r8;
FlatPtr r9;
FlatPtr r10;
FlatPtr r11;
FlatPtr r12;
FlatPtr r13;
FlatPtr r14;
FlatPtr r15;
FlatPtr rip;
FlatPtr rsp0;
#endif
FlatPtr cs;
#if ARCH(I386)
FlatPtr eflags;
FlatPtr flags() const { return eflags; }
void set_flags(FlatPtr value) { eflags = value; }
void set_sp(FlatPtr value) { esp = value; }
void set_sp0(FlatPtr value) { esp0 = value; }
void set_ip(FlatPtr value) { eip = value; }
#else
FlatPtr rflags;
FlatPtr flags() const { return rflags; }
void set_flags(FlatPtr value) { rflags = value; }
void set_sp(FlatPtr value) { rsp = value; }
void set_sp0(FlatPtr value) { rsp0 = value; }
void set_ip(FlatPtr value) { rip = value; }
#endif
FlatPtr cr3;
FlatPtr ip() const
{
#if ARCH(I386)
return eip;
#else
return rip;
#endif
}
FlatPtr sp() const
{
#if ARCH(I386)
return esp;
#else
return rsp;
#endif
}
};
class Thread
: public ListedRefCounted<Thread>
, public Weakable<Thread> {
AK_MAKE_NONCOPYABLE(Thread);
AK_MAKE_NONMOVABLE(Thread);
friend class Mutex;
friend class Process;
friend class Scheduler;
friend struct ThreadReadyQueue;
public:
inline static Thread* current()
{
return Processor::current_thread();
}
static ErrorOr<NonnullRefPtr<Thread>> try_create(NonnullRefPtr<Process>);
~Thread();
static RefPtr<Thread> from_tid(ThreadID);
static void finalize_dying_threads();
ThreadID tid() const { return m_tid; }
ProcessID pid() const;
void set_priority(u32 p) { m_priority = p; }
u32 priority() const { return m_priority; }
void detach()
{
SpinlockLocker lock(m_lock);
m_is_joinable = false;
}
[[nodiscard]] bool is_joinable() const
{
SpinlockLocker lock(m_lock);
return m_is_joinable;
}
Process& process() { return m_process; }
const Process& process() const { return m_process; }
// NOTE: This returns a null-terminated string.
StringView name() const
{
// NOTE: Whoever is calling this needs to be holding our lock while reading the name.
VERIFY(m_lock.is_locked_by_current_processor());
return m_name->view();
}
void set_name(NonnullOwnPtr<KString> name)
{
SpinlockLocker lock(m_lock);
m_name = move(name);
}
void finalize();
enum State : u8 {
Invalid = 0,
Runnable,
Running,
Dying,
Dead,
Stopped,
Blocked
};
class [[nodiscard]] BlockResult {
public:
enum Type {
WokeNormally,
NotBlocked,
InterruptedBySignal,
InterruptedByDeath,
InterruptedByTimeout,
};
BlockResult() = delete;
BlockResult(Type type)
: m_type(type)
{
}
bool operator==(Type type) const
{
return m_type == type;
}
bool operator!=(Type type) const
{
return m_type != type;
}
[[nodiscard]] bool was_interrupted() const
{
switch (m_type) {
case InterruptedBySignal:
case InterruptedByDeath:
return true;
default:
return false;
}
}
private:
Type m_type;
};
class BlockTimeout {
public:
BlockTimeout()
: m_infinite(true)
{
}
explicit BlockTimeout(bool is_absolute, const Time* time, const Time* start_time = nullptr, clockid_t clock_id = CLOCK_MONOTONIC_COARSE);
const Time& absolute_time() const { return m_time; }
const Time* start_time() const { return !m_infinite ? &m_start_time : nullptr; }
clockid_t clock_id() const { return m_clock_id; }
bool is_infinite() const { return m_infinite; }
private:
Time m_time {};
Time m_start_time {};
clockid_t m_clock_id { CLOCK_MONOTONIC_COARSE };
bool m_infinite { false };
};
class BlockerSet;
class Blocker {
AK_MAKE_NONMOVABLE(Blocker);
AK_MAKE_NONCOPYABLE(Blocker);
public:
enum class Type {
Unknown = 0,
File,
Futex,
Plan9FS,
Join,
Queue,
Routing,
Sleep,
Signal,
Wait
};
virtual ~Blocker();
virtual StringView state_string() const = 0;
virtual Type blocker_type() const = 0;
virtual const BlockTimeout& override_timeout(const BlockTimeout& timeout) { return timeout; }
virtual bool can_be_interrupted() const { return true; }
virtual bool setup_blocker();
Thread& thread() { return m_thread; }
enum class UnblockImmediatelyReason {
UnblockConditionAlreadyMet,
TimeoutInThePast,
};
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) = 0;
virtual void was_unblocked(bool did_timeout)
{
if (did_timeout) {
SpinlockLocker lock(m_lock);
m_did_timeout = true;
}
}
void set_interrupted_by_death()
{
SpinlockLocker lock(m_lock);
do_set_interrupted_by_death();
}
void set_interrupted_by_signal(u8 signal)
{
SpinlockLocker lock(m_lock);
do_set_interrupted_by_signal(signal);
}
u8 was_interrupted_by_signal() const
{
SpinlockLocker lock(m_lock);
return do_get_interrupted_by_signal();
}
virtual Thread::BlockResult block_result()
{
SpinlockLocker lock(m_lock);
if (m_was_interrupted_by_death)
return Thread::BlockResult::InterruptedByDeath;
if (m_was_interrupted_by_signal != 0)
return Thread::BlockResult::InterruptedBySignal;
if (m_did_timeout)
return Thread::BlockResult::InterruptedByTimeout;
return Thread::BlockResult::WokeNormally;
}
void begin_blocking(Badge<Thread>);
BlockResult end_blocking(Badge<Thread>, bool);
protected:
Blocker()
: m_thread(*Thread::current())
{
}
void do_set_interrupted_by_death()
{
m_was_interrupted_by_death = true;
}
void do_set_interrupted_by_signal(u8 signal)
{
VERIFY(signal != 0);
m_was_interrupted_by_signal = signal;
}
void do_clear_interrupted_by_signal()
{
m_was_interrupted_by_signal = 0;
}
u8 do_get_interrupted_by_signal() const
{
return m_was_interrupted_by_signal;
}
[[nodiscard]] bool was_interrupted() const
{
return m_was_interrupted_by_death || m_was_interrupted_by_signal != 0;
}
void unblock_from_blocker()
{
{
SpinlockLocker lock(m_lock);
if (!m_is_blocking)
return;
m_is_blocking = false;
}
m_thread->unblock_from_blocker(*this);
}
bool add_to_blocker_set(BlockerSet&, void* = nullptr);
void set_blocker_set_raw_locked(BlockerSet* blocker_set) { m_blocker_set = blocker_set; }
mutable RecursiveSpinlock m_lock;
private:
BlockerSet* m_blocker_set { nullptr };
NonnullRefPtr<Thread> m_thread;
u8 m_was_interrupted_by_signal { 0 };
bool m_is_blocking { false };
bool m_was_interrupted_by_death { false };
bool m_did_timeout { false };
};
class BlockerSet {
AK_MAKE_NONCOPYABLE(BlockerSet);
AK_MAKE_NONMOVABLE(BlockerSet);
public:
BlockerSet() = default;
virtual ~BlockerSet()
{
VERIFY(!m_lock.is_locked());
VERIFY(m_blockers.is_empty());
}
bool add_blocker(Blocker& blocker, void* data)
{
SpinlockLocker lock(m_lock);
if (!should_add_blocker(blocker, data))
return false;
m_blockers.append({ &blocker, data });
return true;
}
void remove_blocker(Blocker& blocker)
{
SpinlockLocker lock(m_lock);
// NOTE: it's possible that the blocker is no longer present
m_blockers.remove_all_matching([&](auto& info) {
return info.blocker == &blocker;
});
}
bool is_empty() const
{
SpinlockLocker lock(m_lock);
return is_empty_locked();
}
protected:
template<typename Callback>
bool unblock_all_blockers_whose_conditions_are_met(Callback try_to_unblock_one)
{
SpinlockLocker lock(m_lock);
return unblock_all_blockers_whose_conditions_are_met_locked(try_to_unblock_one);
}
template<typename Callback>
bool unblock_all_blockers_whose_conditions_are_met_locked(Callback try_to_unblock_one)
{
VERIFY(m_lock.is_locked());
bool stop_iterating = false;
bool did_unblock_any = false;
for (size_t i = 0; i < m_blockers.size() && !stop_iterating;) {
auto& info = m_blockers[i];
if (bool did_unblock = try_to_unblock_one(*info.blocker, info.data, stop_iterating)) {
m_blockers.remove(i);
did_unblock_any = true;
continue;
}
i++;
}
return did_unblock_any;
}
bool is_empty_locked() const
{
VERIFY(m_lock.is_locked());
return m_blockers.is_empty();
}
virtual bool should_add_blocker(Blocker&, void*) { return true; }
struct BlockerInfo {
Blocker* blocker;
void* data;
};
Vector<BlockerInfo, 4> do_take_blockers(size_t count)
{
if (m_blockers.size() <= count)
return move(m_blockers);
size_t move_count = (count <= m_blockers.size()) ? count : m_blockers.size();
VERIFY(move_count > 0);
Vector<BlockerInfo, 4> taken_blockers;
taken_blockers.ensure_capacity(move_count);
for (size_t i = 0; i < move_count; i++)
taken_blockers.append(m_blockers.take(i));
m_blockers.remove(0, move_count);
return taken_blockers;
}
void do_append_blockers(Vector<BlockerInfo, 4>&& blockers_to_append)
{
if (blockers_to_append.is_empty())
return;
if (m_blockers.is_empty()) {
m_blockers = move(blockers_to_append);
return;
}
m_blockers.ensure_capacity(m_blockers.size() + blockers_to_append.size());
for (size_t i = 0; i < blockers_to_append.size(); i++)
m_blockers.append(blockers_to_append.take(i));
blockers_to_append.clear();
}
mutable Spinlock m_lock;
private:
Vector<BlockerInfo, 4> m_blockers;
};
friend class JoinBlocker;
class JoinBlocker final : public Blocker {
public:
explicit JoinBlocker(Thread& joinee, ErrorOr<void>& try_join_result, void*& joinee_exit_value);
virtual Type blocker_type() const override { return Type::Join; }
virtual StringView state_string() const override { return "Joining"sv; }
virtual bool can_be_interrupted() const override { return false; }
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override;
virtual bool setup_blocker() override;
bool unblock(void*, bool);
private:
NonnullRefPtr<Thread> m_joinee;
void*& m_joinee_exit_value;
ErrorOr<void>& m_try_join_result;
bool m_did_unblock { false };
};
class WaitQueueBlocker final : public Blocker {
public:
explicit WaitQueueBlocker(WaitQueue&, StringView block_reason = {});
virtual ~WaitQueueBlocker();
virtual Type blocker_type() const override { return Type::Queue; }
virtual StringView state_string() const override { return m_block_reason.is_null() ? m_block_reason : "Queue"sv; }
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override { }
virtual bool setup_blocker() override;
bool unblock();
protected:
WaitQueue& m_wait_queue;
StringView m_block_reason;
bool m_did_unblock { false };
};
class FutexBlocker final : public Blocker {
public:
explicit FutexBlocker(FutexQueue&, u32);
virtual ~FutexBlocker();
virtual Type blocker_type() const override { return Type::Futex; }
virtual StringView state_string() const override { return "Futex"sv; }
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override { }
virtual bool setup_blocker() override;
u32 bitset() const { return m_bitset; }
void begin_requeue()
{
// We need to hold the lock until we moved it over
m_relock_flags = m_lock.lock();
}
void finish_requeue(FutexQueue&);
bool unblock_bitset(u32 bitset);
bool unblock(bool force = false);
protected:
FutexQueue& m_futex_queue;
u32 m_bitset { 0 };
u32 m_relock_flags { 0 };
bool m_did_unblock { false };
};
class FileBlocker : public Blocker {
public:
enum class BlockFlags : u16 {
None = 0,
Read = 1 << 0,
Write = 1 << 1,
ReadPriority = 1 << 2,
WritePriority = 1 << 3,
Accept = 1 << 4,
Connect = 1 << 5,
SocketFlags = Accept | Connect,
WriteNotOpen = 1 << 6,
WriteError = 1 << 7,
WriteHangUp = 1 << 8,
ReadHangUp = 1 << 9,
Exception = WriteNotOpen | WriteError | WriteHangUp | ReadHangUp,
};
virtual Type blocker_type() const override { return Type::File; }
virtual bool unblock_if_conditions_are_met(bool, void*) = 0;
};
class OpenFileDescriptionBlocker : public FileBlocker {
public:
const OpenFileDescription& blocked_description() const;
virtual bool unblock_if_conditions_are_met(bool, void*) override;
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override;
virtual bool setup_blocker() override;
protected:
explicit OpenFileDescriptionBlocker(OpenFileDescription&, BlockFlags, BlockFlags&);
private:
NonnullRefPtr<OpenFileDescription> m_blocked_description;
const BlockFlags m_flags;
BlockFlags& m_unblocked_flags;
bool m_did_unblock { false };
};
class AcceptBlocker final : public OpenFileDescriptionBlocker {
public:
explicit AcceptBlocker(OpenFileDescription&, BlockFlags&);
virtual StringView state_string() const override { return "Accepting"sv; }
};
class ConnectBlocker final : public OpenFileDescriptionBlocker {
public:
explicit ConnectBlocker(OpenFileDescription&, BlockFlags&);
virtual StringView state_string() const override { return "Connecting"sv; }
};
class WriteBlocker final : public OpenFileDescriptionBlocker {
public:
explicit WriteBlocker(OpenFileDescription&, BlockFlags&);
virtual StringView state_string() const override { return "Writing"sv; }
virtual const BlockTimeout& override_timeout(const BlockTimeout&) override;
private:
BlockTimeout m_timeout;
};
class ReadBlocker final : public OpenFileDescriptionBlocker {
public:
explicit ReadBlocker(OpenFileDescription&, BlockFlags&);
virtual StringView state_string() const override { return "Reading"sv; }
virtual const BlockTimeout& override_timeout(const BlockTimeout&) override;
private:
BlockTimeout m_timeout;
};
class SleepBlocker final : public Blocker {
public:
explicit SleepBlocker(const BlockTimeout&, Time* = nullptr);
virtual StringView state_string() const override { return "Sleeping"sv; }
virtual Type blocker_type() const override { return Type::Sleep; }
virtual const BlockTimeout& override_timeout(const BlockTimeout&) override;
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override;
virtual void was_unblocked(bool) override;
virtual Thread::BlockResult block_result() override;
private:
void calculate_remaining();
BlockTimeout m_deadline;
Time* m_remaining;
};
class SelectBlocker final : public FileBlocker {
public:
struct FDInfo {
NonnullRefPtr<OpenFileDescription> description;
BlockFlags block_flags { BlockFlags::None };
BlockFlags unblocked_flags { BlockFlags::None };
};
using FDVector = Vector<FDInfo, FD_SETSIZE>;
explicit SelectBlocker(FDVector&);
virtual ~SelectBlocker();
virtual bool unblock_if_conditions_are_met(bool, void*) override;
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override;
virtual void was_unblocked(bool) override;
virtual StringView state_string() const override { return "Selecting"sv; }
virtual bool setup_blocker() override;
private:
size_t collect_unblocked_flags();
FDVector& m_fds;
bool m_did_unblock { false };
};
class SignalBlocker final : public Blocker {
public:
explicit SignalBlocker(sigset_t pending_set, siginfo_t& result);
virtual StringView state_string() const override { return "Pending Signal"sv; }
virtual Type blocker_type() const override { return Type::Signal; }
void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override;
virtual bool setup_blocker() override;
bool check_pending_signals(bool from_add_blocker);
private:
sigset_t m_pending_set { 0 };
siginfo_t& m_result;
bool m_did_unblock { false };
};
class SignalBlockerSet final : public BlockerSet {
public:
void unblock_all_blockers_whose_conditions_are_met()
{
BlockerSet::unblock_all_blockers_whose_conditions_are_met([&](auto& b, void*, bool&) {
VERIFY(b.blocker_type() == Blocker::Type::Signal);
auto& blocker = static_cast<Thread::SignalBlocker&>(b);
return blocker.check_pending_signals(false);
});
}
private:
bool should_add_blocker(Blocker& b, void*) override
{
VERIFY(b.blocker_type() == Blocker::Type::Signal);
auto& blocker = static_cast<Thread::SignalBlocker&>(b);
return !blocker.check_pending_signals(true);
}
};
class WaitBlocker final : public Blocker {
public:
enum class UnblockFlags {
Terminated,
Stopped,
Continued,
Disowned
};
WaitBlocker(int wait_options, Variant<Empty, NonnullRefPtr<Process>, NonnullRefPtr<ProcessGroup>> waitee, ErrorOr<siginfo_t>& result);
virtual StringView state_string() const override { return "Waiting"sv; }
virtual Type blocker_type() const override { return Type::Wait; }
virtual void will_unblock_immediately_without_blocking(UnblockImmediatelyReason) override;
virtual void was_unblocked(bool) override;
virtual bool setup_blocker() override;
bool unblock(Process& process, UnblockFlags flags, u8 signal, bool from_add_blocker);
bool is_wait() const { return (m_wait_options & WNOWAIT) != WNOWAIT; }
private:
void do_was_disowned();
void do_set_result(const siginfo_t&);
const int m_wait_options;
ErrorOr<siginfo_t>& m_result;
Variant<Empty, NonnullRefPtr<Process>, NonnullRefPtr<ProcessGroup>> m_waitee;
bool m_did_unblock { false };
bool m_got_sigchild { false };
};
class WaitBlockerSet final : public BlockerSet {
friend class WaitBlocker;
public:
explicit WaitBlockerSet(Process& process)
: m_process(process)
{
}
void disowned_by_waiter(Process&);
bool unblock(Process&, WaitBlocker::UnblockFlags, u8);
void try_unblock(WaitBlocker&);
void finalize();
protected:
virtual bool should_add_blocker(Blocker&, void*) override;
private:
struct ProcessBlockInfo {
NonnullRefPtr<Process> process;
WaitBlocker::UnblockFlags flags;
u8 signal;
bool was_waited { false };
explicit ProcessBlockInfo(NonnullRefPtr<Process>&&, WaitBlocker::UnblockFlags, u8);
~ProcessBlockInfo();
};
Process& m_process;
Vector<ProcessBlockInfo, 2> m_processes;
bool m_finalized { false };
};
template<typename AddBlockerHandler>
ErrorOr<void> try_join(AddBlockerHandler add_blocker)
{
if (Thread::current() == this)
return EDEADLK;
SpinlockLocker lock(m_lock);
if (!m_is_joinable || state() == Dead)
return EINVAL;
add_blocker();
// From this point on the thread is no longer joinable by anyone
// else. It also means that if the join is timed, it becomes
// detached when a timeout happens.
m_is_joinable = false;
return {};
}
void did_schedule() { ++m_times_scheduled; }
u32 times_scheduled() const { return m_times_scheduled; }
void resume_from_stopped();
[[nodiscard]] bool should_be_stopped() const;
[[nodiscard]] bool is_stopped() const { return m_state == Stopped; }
[[nodiscard]] bool is_blocked() const { return m_state == Blocked; }
[[nodiscard]] bool is_in_block() const
{
SpinlockLocker lock(m_block_lock);
return m_in_block;
}
u32 cpu() const { return m_cpu.load(AK::MemoryOrder::memory_order_consume); }
void set_cpu(u32 cpu) { m_cpu.store(cpu, AK::MemoryOrder::memory_order_release); }
u32 affinity() const { return m_cpu_affinity; }
void set_affinity(u32 affinity) { m_cpu_affinity = affinity; }
RegisterState& get_register_dump_from_stack();
const RegisterState& get_register_dump_from_stack() const { return const_cast<Thread*>(this)->get_register_dump_from_stack(); }
DebugRegisterState& debug_register_state() { return m_debug_register_state; }
const DebugRegisterState& debug_register_state() const { return m_debug_register_state; }
ThreadRegisters& regs() { return m_regs; }
ThreadRegisters const& regs() const { return m_regs; }
State state() const { return m_state; }
StringView state_string() const;
VirtualAddress thread_specific_data() const { return m_thread_specific_data; }
size_t thread_specific_region_size() const;
size_t thread_specific_region_alignment() const;
ALWAYS_INLINE void yield_if_stopped()
{
// If some thread stopped us, we need to yield to someone else
// We check this when entering/exiting a system call. A thread
// may continue to execute in user land until the next timer
// tick or entering the next system call, or if it's in kernel
// mode then we will intercept prior to returning back to user
// mode.
SpinlockLocker lock(m_lock);
while (state() == Thread::Stopped) {
lock.unlock();
// We shouldn't be holding the big lock here
yield_without_releasing_big_lock();
lock.lock();
}
}
void block(Kernel::Mutex&, SpinlockLocker<Spinlock>&, u32);
template<typename BlockerType, class... Args>
[[nodiscard]] BlockResult block(const BlockTimeout& timeout, Args&&... args)
{
VERIFY(!Processor::current_in_irq());
VERIFY(this == Thread::current());
ScopedCritical critical;
VERIFY(!Memory::s_mm_lock.is_locked_by_current_processor());
SpinlockLocker block_lock(m_block_lock);
// We need to hold m_block_lock so that nobody can unblock a blocker as soon
// as it is constructed and registered elsewhere
VERIFY(!m_in_block);
TemporaryChange in_block_change(m_in_block, true);
BlockerType blocker(forward<Args>(args)...);
if (!blocker.setup_blocker()) {
blocker.will_unblock_immediately_without_blocking(Blocker::UnblockImmediatelyReason::UnblockConditionAlreadyMet);
return BlockResult::NotBlocked;
}
SpinlockLocker scheduler_lock(g_scheduler_lock);
// Relaxed semantics are fine for timeout_unblocked because we
// synchronize on the spin locks already.
Atomic<bool, AK::MemoryOrder::memory_order_relaxed> timeout_unblocked(false);
bool timer_was_added = false;
switch (state()) {
case Thread::Stopped:
// It's possible that we were requested to be stopped!
break;
case Thread::Running:
VERIFY(m_blocker == nullptr);
break;
default:
VERIFY_NOT_REACHED();
}
m_blocker = &blocker;
if (auto& block_timeout = blocker.override_timeout(timeout); !block_timeout.is_infinite()) {
// Process::kill_all_threads may be called at any time, which will mark all
// threads to die. In that case
timer_was_added = TimerQueue::the().add_timer_without_id(*m_block_timer, block_timeout.clock_id(), block_timeout.absolute_time(), [&]() {
VERIFY(!Processor::current_in_irq());
VERIFY(!g_scheduler_lock.is_locked_by_current_processor());
VERIFY(!m_block_lock.is_locked_by_current_processor());
// NOTE: this may execute on the same or any other processor!
SpinlockLocker scheduler_lock(g_scheduler_lock);
SpinlockLocker block_lock(m_block_lock);
if (m_blocker && !timeout_unblocked.exchange(true))
unblock();
});
if (!timer_was_added) {
// Timeout is already in the past
blocker.will_unblock_immediately_without_blocking(Blocker::UnblockImmediatelyReason::TimeoutInThePast);
m_blocker = nullptr;
return BlockResult::InterruptedByTimeout;
}
}
blocker.begin_blocking({});
set_state(Thread::Blocked);
scheduler_lock.unlock();
block_lock.unlock();
dbgln_if(THREAD_DEBUG, "Thread {} blocking on {} ({}) -->", *this, &blocker, blocker.state_string());
bool did_timeout = false;
u32 lock_count_to_restore = 0;
auto previous_locked = unlock_process_if_locked(lock_count_to_restore);
for (;;) {
// Yield to the scheduler, and wait for us to resume unblocked.
VERIFY(!g_scheduler_lock.is_locked_by_current_processor());
VERIFY(Processor::in_critical());
yield_without_releasing_big_lock();
VERIFY(Processor::in_critical());
SpinlockLocker block_lock2(m_block_lock);
if (should_be_stopped() || state() == Stopped) {
dbgln("Thread should be stopped, current state: {}", state_string());
set_state(Thread::Blocked);
continue;
}
if (m_blocker && !m_blocker->can_be_interrupted() && !m_should_die) {
block_lock2.unlock();
dbgln("Thread should not be unblocking, current state: {}", state_string());
set_state(Thread::Blocked);
continue;
}
// Prevent the timeout from unblocking this thread if it happens to
// be in the process of firing already
did_timeout |= timeout_unblocked.exchange(true);
if (m_blocker) {
// Remove ourselves...
VERIFY(m_blocker == &blocker);
m_blocker = nullptr;
}
dbgln_if(THREAD_DEBUG, "<-- Thread {} unblocked from {} ({})", *this, &blocker, blocker.state_string());
break;
}
if (blocker.was_interrupted_by_signal()) {
SpinlockLocker scheduler_lock(g_scheduler_lock);
SpinlockLocker lock(m_lock);
dispatch_one_pending_signal();
}
// Notify the blocker that we are no longer blocking. It may need
// to clean up now while we're still holding m_lock
auto result = blocker.end_blocking({}, did_timeout); // calls was_unblocked internally
if (timer_was_added && !did_timeout) {
// Cancel the timer while not holding any locks. This allows
// the timer function to complete before we remove it
// (e.g. if it's on another processor)
TimerQueue::the().cancel_timer(*m_block_timer);
}
if (previous_locked != LockMode::Unlocked) {
// NOTE: this may trigger another call to Thread::block(), so
// we need to do this after we're all done and restored m_in_block!
relock_process(previous_locked, lock_count_to_restore);
}
return result;
}
u32 unblock_from_lock(Kernel::Mutex&);
void unblock_from_blocker(Blocker&);
void unblock(u8 signal = 0);
template<class... Args>
Thread::BlockResult wait_on(WaitQueue& wait_queue, const Thread::BlockTimeout& timeout, Args&&... args)
{
VERIFY(this == Thread::current());
return block<Thread::WaitQueueBlocker>(timeout, wait_queue, forward<Args>(args)...);
}
BlockResult sleep(clockid_t, const Time&, Time* = nullptr);
BlockResult sleep(const Time& duration, Time* remaining_time = nullptr)
{
return sleep(CLOCK_MONOTONIC_COARSE, duration, remaining_time);
}
BlockResult sleep_until(clockid_t, const Time&);
BlockResult sleep_until(const Time& duration)
{
return sleep_until(CLOCK_MONOTONIC_COARSE, duration);
}
// Tell this thread to unblock if needed,
// gracefully unwind the stack and die.
void set_should_die();
[[nodiscard]] bool should_die() const { return m_should_die; }
void die_if_needed();
void exit(void* = nullptr);
void update_time_scheduled(u64, bool, bool);
bool tick();
void set_ticks_left(u32 t) { m_ticks_left = t; }
u32 ticks_left() const { return m_ticks_left; }
FlatPtr kernel_stack_base() const { return m_kernel_stack_base; }
FlatPtr kernel_stack_top() const { return m_kernel_stack_top; }
void set_state(State, u8 = 0);
[[nodiscard]] bool is_initialized() const { return m_initialized; }
void set_initialized(bool initialized) { m_initialized = initialized; }
void send_urgent_signal_to_self(u8 signal);
void send_signal(u8 signal, Process* sender);
u32 update_signal_mask(u32 signal_mask);
u32 signal_mask_block(sigset_t signal_set, bool block);
u32 signal_mask() const;
void reset_signals_for_exec();
ErrorOr<FlatPtr> peek_debug_register(u32 register_index);
ErrorOr<void> poke_debug_register(u32 register_index, FlatPtr data);
void set_dump_backtrace_on_finalization() { m_dump_backtrace_on_finalization = true; }
DispatchSignalResult dispatch_one_pending_signal();
DispatchSignalResult try_dispatch_one_pending_signal(u8 signal);
DispatchSignalResult dispatch_signal(u8 signal);
void check_dispatch_pending_signal();
[[nodiscard]] bool has_unmasked_pending_signals() const { return m_have_any_unmasked_pending_signals.load(AK::memory_order_consume); }
[[nodiscard]] bool should_ignore_signal(u8 signal) const;
[[nodiscard]] bool has_signal_handler(u8 signal) const;
[[nodiscard]] bool is_signal_masked(u8 signal) const;
u32 pending_signals() const;
u32 pending_signals_for_state() const;
[[nodiscard]] bool has_alternative_signal_stack() const;
[[nodiscard]] bool is_in_alternative_signal_stack() const;
FPUState& fpu_state() { return m_fpu_state; }
ErrorOr<void> make_thread_specific_region(Badge<Process>);
unsigned syscall_count() const { return m_syscall_count; }
void did_syscall() { ++m_syscall_count; }
unsigned inode_faults() const { return m_inode_faults; }
void did_inode_fault() { ++m_inode_faults; }
unsigned zero_faults() const { return m_zero_faults; }
void did_zero_fault() { ++m_zero_faults; }
unsigned cow_faults() const { return m_cow_faults; }
void did_cow_fault() { ++m_cow_faults; }
unsigned file_read_bytes() const { return m_file_read_bytes; }
unsigned file_write_bytes() const { return m_file_write_bytes; }
void did_file_read(unsigned bytes)
{
m_file_read_bytes += bytes;
}
void did_file_write(unsigned bytes)
{
m_file_write_bytes += bytes;
}
unsigned unix_socket_read_bytes() const { return m_unix_socket_read_bytes; }
unsigned unix_socket_write_bytes() const { return m_unix_socket_write_bytes; }
void did_unix_socket_read(unsigned bytes)
{
m_unix_socket_read_bytes += bytes;
}
void did_unix_socket_write(unsigned bytes)
{
m_unix_socket_write_bytes += bytes;
}
unsigned ipv4_socket_read_bytes() const { return m_ipv4_socket_read_bytes; }
unsigned ipv4_socket_write_bytes() const { return m_ipv4_socket_write_bytes; }
void did_ipv4_socket_read(unsigned bytes)
{
m_ipv4_socket_read_bytes += bytes;
}
void did_ipv4_socket_write(unsigned bytes)
{
m_ipv4_socket_write_bytes += bytes;
}
void set_active(bool active) { m_is_active = active; }
u32 saved_critical() const { return m_saved_critical; }
void save_critical(u32 critical) { m_saved_critical = critical; }
void track_lock_acquire(LockRank rank);
void track_lock_release(LockRank rank);
[[nodiscard]] bool is_active() const { return m_is_active; }
[[nodiscard]] bool is_finalizable() const
{
// We can't finalize as long as this thread is still running
// Note that checking for Running state here isn't sufficient
// as the thread may not be in Running state but switching out.
// m_is_active is set to false once the context switch is
// complete and the thread is not executing on any processor.
if (m_is_active.load(AK::memory_order_acquire))
return false;
// We can't finalize until the thread is either detached or
// a join has started. We can't make m_is_joinable atomic
// because that would introduce a race in try_join.
SpinlockLocker lock(m_lock);
return !m_is_joinable;
}
ErrorOr<NonnullRefPtr<Thread>> try_clone(Process&);
template<IteratorFunction<Thread&> Callback>
static IterationDecision for_each_in_state(State, Callback);
template<IteratorFunction<Thread&> Callback>
static IterationDecision for_each(Callback);
template<VoidFunction<Thread&> Callback>
static IterationDecision for_each_in_state(State, Callback);
template<VoidFunction<Thread&> Callback>
static IterationDecision for_each(Callback);
static constexpr u32 default_kernel_stack_size = 65536;
static constexpr u32 default_userspace_stack_size = 1 * MiB;
u64 time_in_user() const { return m_total_time_scheduled_user; }
u64 time_in_kernel() const { return m_total_time_scheduled_kernel; }
enum class PreviousMode : u8 {
KernelMode = 0,
UserMode
};
PreviousMode previous_mode() const { return m_previous_mode; }
bool set_previous_mode(PreviousMode mode)
{
if (m_previous_mode == mode)
return false;
m_previous_mode = mode;
return true;
}
TrapFrame*& current_trap() { return m_current_trap; }
TrapFrame const* const& current_trap() const { return m_current_trap; }
RecursiveSpinlock& get_lock() const { return m_lock; }
#if LOCK_DEBUG
void holding_lock(Mutex& lock, int refs_delta, LockLocation const& location)
{
VERIFY(refs_delta != 0);
m_holding_locks.fetch_add(refs_delta, AK::MemoryOrder::memory_order_relaxed);
SpinlockLocker list_lock(m_holding_locks_lock);
if (refs_delta > 0) {
bool have_existing = false;
for (size_t i = 0; i < m_holding_locks_list.size(); i++) {
auto& info = m_holding_locks_list[i];
if (info.lock == &lock) {
have_existing = true;
info.count += refs_delta;
break;
}
}
if (!have_existing)
m_holding_locks_list.append({ &lock, location, 1 });
} else {
VERIFY(refs_delta < 0);
bool found = false;
for (size_t i = 0; i < m_holding_locks_list.size(); i++) {
auto& info = m_holding_locks_list[i];
if (info.lock == &lock) {
VERIFY(info.count >= (unsigned)-refs_delta);
info.count -= (unsigned)-refs_delta;
if (info.count == 0)
m_holding_locks_list.remove(i);
found = true;
break;
}
}
VERIFY(found);
}
}
u32 lock_count() const
{
return m_holding_locks.load(AK::MemoryOrder::memory_order_relaxed);
}
#endif
bool is_handling_page_fault() const
{
return m_handling_page_fault;
}
void set_handling_page_fault(bool b) { m_handling_page_fault = b; }
void set_idle_thread() { m_is_idle_thread = true; }
bool is_idle_thread() const { return m_is_idle_thread; }
void set_crashing() { m_is_crashing = true; }
[[nodiscard]] bool is_crashing() const { return m_is_crashing; }
ALWAYS_INLINE u32 enter_profiler()
{
return m_nested_profiler_calls.fetch_add(1, AK::MemoryOrder::memory_order_acq_rel);
}
ALWAYS_INLINE u32 leave_profiler()
{
return m_nested_profiler_calls.fetch_sub(1, AK::MemoryOrder::memory_order_acquire);
}
bool is_profiling_suppressed() const { return m_is_profiling_suppressed; }
void set_profiling_suppressed() { m_is_profiling_suppressed = true; }
String backtrace();
private:
Thread(NonnullRefPtr<Process>, NonnullOwnPtr<Memory::Region>, NonnullRefPtr<Timer>, NonnullOwnPtr<KString>);
IntrusiveListNode<Thread> m_process_thread_list_node;
int m_runnable_priority { -1 };
friend class WaitQueue;
class JoinBlockerSet final : public BlockerSet {
public:
void thread_did_exit(void* exit_value)
{
SpinlockLocker lock(m_lock);
VERIFY(!m_thread_did_exit);
m_thread_did_exit = true;
m_exit_value.store(exit_value, AK::MemoryOrder::memory_order_release);
do_unblock_joiner();
}
void thread_finalizing()
{
SpinlockLocker lock(m_lock);
do_unblock_joiner();
}
void* exit_value() const
{
VERIFY(m_thread_did_exit);
return m_exit_value.load(AK::MemoryOrder::memory_order_acquire);
}
void try_unblock(JoinBlocker& blocker)
{
SpinlockLocker lock(m_lock);
if (m_thread_did_exit)
blocker.unblock(exit_value(), false);
}
protected:
virtual bool should_add_blocker(Blocker& b, void*) override
{
VERIFY(b.blocker_type() == Blocker::Type::Join);
auto& blocker = static_cast<JoinBlocker&>(b);
// NOTE: m_lock is held already!
if (m_thread_did_exit) {
blocker.unblock(exit_value(), true);
return false;
}
return true;
}
private:
void do_unblock_joiner()
{
unblock_all_blockers_whose_conditions_are_met_locked([&](Blocker& b, void*, bool&) {
VERIFY(b.blocker_type() == Blocker::Type::Join);
auto& blocker = static_cast<JoinBlocker&>(b);
return blocker.unblock(exit_value(), false);
});
}
Atomic<void*> m_exit_value { nullptr };
bool m_thread_did_exit { false };
};
LockMode unlock_process_if_locked(u32&);
void relock_process(LockMode, u32);
void reset_fpu_state();
mutable RecursiveSpinlock m_lock { LockRank::Thread };
mutable RecursiveSpinlock m_block_lock;
NonnullRefPtr<Process> m_process;
ThreadID m_tid { -1 };
ThreadRegisters m_regs {};
DebugRegisterState m_debug_register_state {};
TrapFrame* m_current_trap { nullptr };
u32 m_saved_critical { 1 };
IntrusiveListNode<Thread> m_ready_queue_node;
Atomic<u32> m_cpu { 0 };
u32 m_cpu_affinity { THREAD_AFFINITY_DEFAULT };
Optional<u64> m_last_time_scheduled;
u64 m_total_time_scheduled_user { 0 };
u64 m_total_time_scheduled_kernel { 0 };
u32 m_ticks_left { 0 };
u32 m_times_scheduled { 0 };
u32 m_ticks_in_user { 0 };
u32 m_ticks_in_kernel { 0 };
u32 m_pending_signals { 0 };
u32 m_signal_mask { 0 };
FlatPtr m_alternative_signal_stack { 0 };
FlatPtr m_alternative_signal_stack_size { 0 };
SignalBlockerSet m_signal_blocker_set;
FlatPtr m_kernel_stack_base { 0 };
FlatPtr m_kernel_stack_top { 0 };
OwnPtr<Memory::Region> m_kernel_stack_region;
VirtualAddress m_thread_specific_data;
Optional<Memory::VirtualRange> m_thread_specific_range;
Array<SignalActionData, NSIG> m_signal_action_data;
Blocker* m_blocker { nullptr };
Kernel::Mutex* m_blocking_lock { nullptr };
u32 m_lock_requested_count { 0 };
IntrusiveListNode<Thread> m_blocked_threads_list_node;
LockRank m_lock_rank_mask { LockRank::None };
#if LOCK_DEBUG
struct HoldingLockInfo {
Mutex* lock;
LockLocation lock_location;
unsigned count;
};
Atomic<u32> m_holding_locks { 0 };
Spinlock m_holding_locks_lock;
Vector<HoldingLockInfo> m_holding_locks_list;
#endif
JoinBlockerSet m_join_blocker_set;
Atomic<bool, AK::MemoryOrder::memory_order_relaxed> m_is_active { false };
bool m_is_joinable { true };
bool m_handling_page_fault { false };
PreviousMode m_previous_mode { PreviousMode::KernelMode }; // We always start out in kernel mode
unsigned m_syscall_count { 0 };
unsigned m_inode_faults { 0 };
unsigned m_zero_faults { 0 };
unsigned m_cow_faults { 0 };
unsigned m_file_read_bytes { 0 };
unsigned m_file_write_bytes { 0 };
unsigned m_unix_socket_read_bytes { 0 };
unsigned m_unix_socket_write_bytes { 0 };
unsigned m_ipv4_socket_read_bytes { 0 };
unsigned m_ipv4_socket_write_bytes { 0 };
FPUState m_fpu_state {};
State m_state { Invalid };
NonnullOwnPtr<KString> m_name;
u32 m_priority { THREAD_PRIORITY_NORMAL };
State m_stop_state { Invalid };
bool m_dump_backtrace_on_finalization { false };
bool m_should_die { false };
bool m_initialized { false };
bool m_in_block { false };
bool m_is_idle_thread { false };
bool m_is_crashing { false };
Atomic<bool> m_have_any_unmasked_pending_signals { false };
Atomic<u32> m_nested_profiler_calls { 0 };
NonnullRefPtr<Timer> m_block_timer;
bool m_is_profiling_suppressed { false };
void yield_and_release_relock_big_lock();
enum class VerifyLockNotHeld {
Yes,
No
};
void yield_without_releasing_big_lock(VerifyLockNotHeld verify_lock_not_held = VerifyLockNotHeld::Yes);
void drop_thread_count(bool);
mutable IntrusiveListNode<Thread> m_global_thread_list_node;
public:
using ListInProcess = IntrusiveList<&Thread::m_process_thread_list_node>;
using GlobalList = IntrusiveList<&Thread::m_global_thread_list_node>;
static SpinlockProtected<GlobalList>& all_instances();
};
AK_ENUM_BITWISE_OPERATORS(Thread::FileBlocker::BlockFlags);
template<IteratorFunction<Thread&> Callback>
inline IterationDecision Thread::for_each(Callback callback)
{
return Thread::all_instances().with([&](auto& list) -> IterationDecision {
for (auto& thread : list) {
IterationDecision decision = callback(thread);
if (decision != IterationDecision::Continue)
return decision;
}
return IterationDecision::Continue;
});
}
template<IteratorFunction<Thread&> Callback>
inline IterationDecision Thread::for_each_in_state(State state, Callback callback)
{
return Thread::all_instances().with([&](auto& list) -> IterationDecision {
for (auto& thread : list) {
if (thread.state() != state)
continue;
IterationDecision decision = callback(thread);
if (decision != IterationDecision::Continue)
return decision;
}
return IterationDecision::Continue;
});
}
template<VoidFunction<Thread&> Callback>
inline IterationDecision Thread::for_each(Callback callback)
{
return Thread::all_instances().with([&](auto& list) {
for (auto& thread : list) {
if (callback(thread) == IterationDecision::Break)
return IterationDecision::Break;
}
return IterationDecision::Continue;
});
}
template<VoidFunction<Thread&> Callback>
inline IterationDecision Thread::for_each_in_state(State state, Callback callback)
{
return for_each_in_state(state, [&](auto& thread) {
callback(thread);
return IterationDecision::Continue;
});
}
}
template<>
struct AK::Formatter<Kernel::Thread> : AK::Formatter<FormatString> {
ErrorOr<void> format(FormatBuilder&, Kernel::Thread const&);
};