emu-mirror/yuzu
1
0
Fork 0
mirror of https://github.com/yuzu-mirror/yuzu.git synced 2024-11-19 16:27:33 +01:00
Code Issues Packages Projects Releases Wiki Activity
b77f571d20
yuzu/src/core/hle/kernel/kernel.cpp

321 lines
11 KiB
C++
Raw Normal View History

License change
2014-12-17 06:38:14 +01:00
// Copyright 2014 Citra Emulator Project
// Licensed under GPLv2 or any later version
Remove trailing spaces in every file but the ones imported from SkyEye, AOSP or generated
2014-11-19 09:49:13 +01:00
// Refer to the license.txt file included.
added initial kernel/thread modules
2014-05-10 04:11:18 +02:00
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
#include <array>
#include <atomic>
#include <memory>
#include <mutex>
#include <utility>
#include "common/assert.h"
#include "common/logging/log.h"
#include "core/core.h"
#include "core/core_timing.h"
service: Migrate global named port map to the KernelCore class Now that we have a class representing the kernel in some capacity, we now have a place to put the named port map, so we move it over and get rid of another piece of global state within the core.
2018-09-02 17:58:58 +02:00
#include "core/hle/kernel/client_port.h"
Kernel: Move HandleTable to a separate file
2017-05-30 01:45:42 +02:00
#include "core/hle/kernel/handle_table.h"
Use negative priorities to avoid special-casing the self-include
2016-09-21 08:52:38 +02:00
#include "core/hle/kernel/kernel.h"
Kernel: Introduce skeleton Process class to hold process data
2015-05-04 05:01:16 +02:00
#include "core/hle/kernel/process.h"
Kernel: Add more infrastructure to support different memory layouts This adds some structures necessary to support multiple memory regions in the future. It also adds support for different system memory types and the new linear heap mapping at 0x30000000.
2015-08-06 02:26:52 +02:00
#include "core/hle/kernel/resource_limit.h"
added initial kernel/thread modules
2014-05-10 04:11:18 +02:00
#include "core/hle/kernel/thread.h"
SVC: Implemented the Timer service calls.
2014-12-04 20:45:47 +01:00
#include "core/hle/kernel/timer.h"
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
#include "core/hle/lock.h"
#include "core/hle/result.h"
added initial kernel/thread modules
2014-05-10 04:11:18 +02:00
- created a Kernel namespace - cleaned up Kernel code a bit (moved stuff into namespace, fixed whitespace issues) - added handle types for all different CTROS handles
2014-05-21 00:13:25 +02:00
namespace Kernel {
added initial kernel/thread modules
2014-05-10 04:11:18 +02:00
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
/**
* Callback that will wake up the thread it was scheduled for
* @param thread_handle The handle of the thread that's been awoken
* @param cycles_late The number of CPU cycles that have passed since the desired wakeup time
*/
static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_late) {
const auto proper_handle = static_cast<Handle>(thread_handle);
auto& system = Core::System::GetInstance();
// Lock the global kernel mutex when we enter the kernel HLE.
std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
SharedPtr<Thread> thread =
system.Kernel().RetrieveThreadFromWakeupCallbackHandleTable(proper_handle);
if (thread == nullptr) {
LOG_CRITICAL(Kernel, "Callback fired for invalid thread {:08X}", proper_handle);
return;
}
bool resume = true;
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
if (thread->GetStatus() == ThreadStatus::WaitSynchAny ||
thread->GetStatus() == ThreadStatus::WaitSynchAll ||
thread->GetStatus() == ThreadStatus::WaitHLEEvent) {
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
// Remove the thread from each of its waiting objects' waitlists
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
for (const auto& object : thread->GetWaitObjects()) {
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
object->RemoveWaitingThread(thread.get());
}
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
thread->ClearWaitObjects();
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
// Invoke the wakeup callback before clearing the wait objects
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
if (thread->HasWakeupCallback()) {
resume = thread->InvokeWakeupCallback(ThreadWakeupReason::Timeout, thread, nullptr, 0);
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
}
}
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
if (thread->GetMutexWaitAddress() != 0 || thread->GetCondVarWaitAddress() != 0 ||
thread->GetWaitHandle() != 0) {
ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex);
thread->SetMutexWaitAddress(0);
thread->SetCondVarWaitAddress(0);
thread->SetWaitHandle(0);
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
auto* const lock_owner = thread->GetLockOwner();
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
// Threads waking up by timeout from WaitProcessWideKey do not perform priority inheritance
// and don't have a lock owner unless SignalProcessWideKey was called first and the thread
// wasn't awakened due to the mutex already being acquired.
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
if (lock_owner != nullptr) {
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
lock_owner->RemoveMutexWaiter(thread);
}
}
kernel/thread: Make all instance variables private Many of the member variables of the thread class aren't even used outside of the class itself, so there's no need to make those variables public. This change follows in the steps of the previous changes that made other kernel types' members private. The main motivation behind this is that the Thread class will likely change in the future as emulation becomes more accurate, and letting random bits of the emulator access data members of the Thread class directly makes it a pain to shuffle around and/or modify internals. Having all data members public like this also makes it difficult to reason about certain bits of behavior without first verifying what parts of the core actually use them. Everything being public also generally follows the tendency for changes to be introduced in completely different translation units that would otherwise be better introduced as an addition to the Thread class' public interface.
2018-10-04 00:47:57 +02:00
if (thread->GetArbiterWaitAddress() != 0) {
ASSERT(thread->GetStatus() == ThreadStatus::WaitArb);
thread->SetArbiterWaitAddress(0);
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
}
if (resume) {
thread->ResumeFromWait();
}
}
/// The timer callback event, called when a timer is fired
static void TimerCallback(u64 timer_handle, int cycles_late) {
const auto proper_handle = static_cast<Handle>(timer_handle);
auto& system = Core::System::GetInstance();
SharedPtr<Timer> timer = system.Kernel().RetrieveTimerFromCallbackHandleTable(proper_handle);
if (timer == nullptr) {
LOG_CRITICAL(Kernel, "Callback fired for invalid timer {:016X}", timer_handle);
return;
}
timer->Signal(cycles_late);
}
- added __KernelLoadExec function - fixed some logging
2014-05-14 03:57:12 +02:00
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
struct KernelCore::Impl {
void Initialize(KernelCore& kernel) {
Shutdown();
Kernel: Properly initialize and shutdown all modules.
2015-04-28 04:12:35 +02:00
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
InitializeResourceLimits(kernel);
InitializeThreads();
InitializeTimers();
}
void Shutdown() {
next_object_id = 0;
next_process_id = 10;
next_thread_id = 1;
process_list.clear();
kernel/thread: Use a regular pointer for the owner/current process There's no real need to use a shared pointer in these cases, and only makes object management more fragile in terms of how easy it would be to introduce cycles. Instead, just do the simple thing of using a regular pointer. Much of this is just a hold-over from citra anyways. It also doesn't make sense from a behavioral point of view for a process' thread to prolong the lifetime of the process itself (the process is supposed to own the thread, not the other way around).
2018-10-10 06:42:10 +02:00
current_process = nullptr;
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
resource_limits.fill(nullptr);
thread_wakeup_callback_handle_table.Clear();
thread_wakeup_event_type = nullptr;
timer_callback_handle_table.Clear();
timer_callback_event_type = nullptr;
service: Migrate global named port map to the KernelCore class Now that we have a class representing the kernel in some capacity, we now have a place to put the named port map, so we move it over and get rid of another piece of global state within the core.
2018-09-02 17:58:58 +02:00
named_ports.clear();
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
}
void InitializeResourceLimits(KernelCore& kernel) {
// Create the four resource limits that the system uses
// Create the APPLICATION resource limit
SharedPtr<ResourceLimit> resource_limit = ResourceLimit::Create(kernel, "Applications");
resource_limit->max_priority = 0x18;
resource_limit->max_commit = 0x4000000;
resource_limit->max_threads = 0x20;
resource_limit->max_events = 0x20;
resource_limit->max_mutexes = 0x20;
resource_limit->max_semaphores = 0x8;
resource_limit->max_timers = 0x8;
resource_limit->max_shared_mems = 0x10;
resource_limit->max_address_arbiters = 0x2;
resource_limit->max_cpu_time = 0x1E;
resource_limits[static_cast<u8>(ResourceLimitCategory::APPLICATION)] = resource_limit;
// Create the SYS_APPLET resource limit
resource_limit = ResourceLimit::Create(kernel, "System Applets");
resource_limit->max_priority = 0x4;
resource_limit->max_commit = 0x5E00000;
resource_limit->max_threads = 0x1D;
resource_limit->max_events = 0xB;
resource_limit->max_mutexes = 0x8;
resource_limit->max_semaphores = 0x4;
resource_limit->max_timers = 0x4;
resource_limit->max_shared_mems = 0x8;
resource_limit->max_address_arbiters = 0x3;
resource_limit->max_cpu_time = 0x2710;
resource_limits[static_cast<u8>(ResourceLimitCategory::SYS_APPLET)] = resource_limit;
// Create the LIB_APPLET resource limit
resource_limit = ResourceLimit::Create(kernel, "Library Applets");
resource_limit->max_priority = 0x4;
resource_limit->max_commit = 0x600000;
resource_limit->max_threads = 0xE;
resource_limit->max_events = 0x8;
resource_limit->max_mutexes = 0x8;
resource_limit->max_semaphores = 0x4;
resource_limit->max_timers = 0x4;
resource_limit->max_shared_mems = 0x8;
resource_limit->max_address_arbiters = 0x1;
resource_limit->max_cpu_time = 0x2710;
resource_limits[static_cast<u8>(ResourceLimitCategory::LIB_APPLET)] = resource_limit;
// Create the OTHER resource limit
resource_limit = ResourceLimit::Create(kernel, "Others");
resource_limit->max_priority = 0x4;
resource_limit->max_commit = 0x2180000;
resource_limit->max_threads = 0xE1;
resource_limit->max_events = 0x108;
resource_limit->max_mutexes = 0x25;
resource_limit->max_semaphores = 0x43;
resource_limit->max_timers = 0x2C;
resource_limit->max_shared_mems = 0x1F;
resource_limit->max_address_arbiters = 0x2D;
resource_limit->max_cpu_time = 0x3E8;
resource_limits[static_cast<u8>(ResourceLimitCategory::OTHER)] = resource_limit;
}
void InitializeThreads() {
thread_wakeup_event_type =
CoreTiming::RegisterEvent("ThreadWakeupCallback", ThreadWakeupCallback);
}
void InitializeTimers() {
timer_callback_handle_table.Clear();
timer_callback_event_type = CoreTiming::RegisterEvent("TimerCallback", TimerCallback);
}
std::atomic<u32> next_object_id{0};
fixup!
2015-05-12 01:23:45 +02:00
// TODO(Subv): Start the process ids from 10 for now, as lower PIDs are
// reserved for low-level services
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
std::atomic<u32> next_process_id{10};
std::atomic<u32> next_thread_id{1};
// Lists all processes that exist in the current session.
std::vector<SharedPtr<Process>> process_list;
kernel/thread: Use a regular pointer for the owner/current process There's no real need to use a shared pointer in these cases, and only makes object management more fragile in terms of how easy it would be to introduce cycles. Instead, just do the simple thing of using a regular pointer. Much of this is just a hold-over from citra anyways. It also doesn't make sense from a behavioral point of view for a process' thread to prolong the lifetime of the process itself (the process is supposed to own the thread, not the other way around).
2018-10-10 06:42:10 +02:00
Process* current_process = nullptr;
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
std::array<SharedPtr<ResourceLimit>, 4> resource_limits;
/// The event type of the generic timer callback event
CoreTiming::EventType* timer_callback_event_type = nullptr;
// TODO(yuriks): This can be removed if Timer objects are explicitly pooled in the future,
// allowing us to simply use a pool index or similar.
Kernel::HandleTable timer_callback_handle_table;
CoreTiming::EventType* thread_wakeup_event_type = nullptr;
// TODO(yuriks): This can be removed if Thread objects are explicitly pooled in the future,
// allowing us to simply use a pool index or similar.
Kernel::HandleTable thread_wakeup_callback_handle_table;
service: Migrate global named port map to the KernelCore class Now that we have a class representing the kernel in some capacity, we now have a place to put the named port map, so we move it over and get rid of another piece of global state within the core.
2018-09-02 17:58:58 +02:00
/// Map of named ports managed by the kernel, which can be retrieved using
/// the ConnectToPort SVC.
NamedPortTable named_ports;
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
};
KernelCore::KernelCore() : impl{std::make_unique<Impl>()} {}
KernelCore::~KernelCore() {
Shutdown();
}
void KernelCore::Initialize() {
impl->Initialize(*this);
}
void KernelCore::Shutdown() {
impl->Shutdown();
}
SharedPtr<ResourceLimit> KernelCore::ResourceLimitForCategory(
ResourceLimitCategory category) const {
return impl->resource_limits.at(static_cast<std::size_t>(category));
}
SharedPtr<Thread> KernelCore::RetrieveThreadFromWakeupCallbackHandleTable(Handle handle) const {
return impl->thread_wakeup_callback_handle_table.Get<Thread>(handle);
}
SharedPtr<Timer> KernelCore::RetrieveTimerFromCallbackHandleTable(Handle handle) const {
return impl->timer_callback_handle_table.Get<Timer>(handle);
}
void KernelCore::AppendNewProcess(SharedPtr<Process> process) {
impl->process_list.push_back(std::move(process));
}
kernel/thread: Use a regular pointer for the owner/current process There's no real need to use a shared pointer in these cases, and only makes object management more fragile in terms of how easy it would be to introduce cycles. Instead, just do the simple thing of using a regular pointer. Much of this is just a hold-over from citra anyways. It also doesn't make sense from a behavioral point of view for a process' thread to prolong the lifetime of the process itself (the process is supposed to own the thread, not the other way around).
2018-10-10 06:42:10 +02:00
void KernelCore::MakeCurrentProcess(Process* process) {
impl->current_process = process;
core: Migrate current_process pointer to the kernel Given we now have the kernel as a class, it doesn't make sense to keep the current process pointer within the System class, as processes are related to the kernel. This also gets rid of a subtle case where memory wouldn't be freed on core shutdown, as the current_process pointer would never be reset, causing the pointed to contents to continue to live.
2018-09-07 02:34:51 +02:00
}
kernel/thread: Use a regular pointer for the owner/current process There's no real need to use a shared pointer in these cases, and only makes object management more fragile in terms of how easy it would be to introduce cycles. Instead, just do the simple thing of using a regular pointer. Much of this is just a hold-over from citra anyways. It also doesn't make sense from a behavioral point of view for a process' thread to prolong the lifetime of the process itself (the process is supposed to own the thread, not the other way around).
2018-10-10 06:42:10 +02:00
Process* KernelCore::CurrentProcess() {
core: Migrate current_process pointer to the kernel Given we now have the kernel as a class, it doesn't make sense to keep the current process pointer within the System class, as processes are related to the kernel. This also gets rid of a subtle case where memory wouldn't be freed on core shutdown, as the current_process pointer would never be reset, causing the pointed to contents to continue to live.
2018-09-07 02:34:51 +02:00
return impl->current_process;
}
kernel/thread: Use a regular pointer for the owner/current process There's no real need to use a shared pointer in these cases, and only makes object management more fragile in terms of how easy it would be to introduce cycles. Instead, just do the simple thing of using a regular pointer. Much of this is just a hold-over from citra anyways. It also doesn't make sense from a behavioral point of view for a process' thread to prolong the lifetime of the process itself (the process is supposed to own the thread, not the other way around).
2018-10-10 06:42:10 +02:00
const Process* KernelCore::CurrentProcess() const {
core: Migrate current_process pointer to the kernel Given we now have the kernel as a class, it doesn't make sense to keep the current process pointer within the System class, as processes are related to the kernel. This also gets rid of a subtle case where memory wouldn't be freed on core shutdown, as the current_process pointer would never be reset, causing the pointed to contents to continue to live.
2018-09-07 02:34:51 +02:00
return impl->current_process;
}
service: Migrate global named port map to the KernelCore class Now that we have a class representing the kernel in some capacity, we now have a place to put the named port map, so we move it over and get rid of another piece of global state within the core.
2018-09-02 17:58:58 +02:00
void KernelCore::AddNamedPort(std::string name, SharedPtr<ClientPort> port) {
impl->named_ports.emplace(std::move(name), std::move(port));
}
KernelCore::NamedPortTable::iterator KernelCore::FindNamedPort(const std::string& name) {
return impl->named_ports.find(name);
}
KernelCore::NamedPortTable::const_iterator KernelCore::FindNamedPort(
const std::string& name) const {
return impl->named_ports.find(name);
}
bool KernelCore::IsValidNamedPort(NamedPortTable::const_iterator port) const {
return port != impl->named_ports.cend();
}
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
u32 KernelCore::CreateNewObjectID() {
return impl->next_object_id++;
}
u32 KernelCore::CreateNewThreadID() {
return impl->next_thread_id++;
- created a Kernel namespace - cleaned up Kernel code a bit (moved stuff into namespace, fixed whitespace issues) - added handle types for all different CTROS handles
2014-05-21 00:13:25 +02:00
}
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
u32 KernelCore::CreateNewProcessID() {
return impl->next_process_id++;
}
Kernel: Add more infrastructure to support different memory layouts This adds some structures necessary to support multiple memory regions in the future. It also adds support for different system memory types and the new linear heap mapping at 0x30000000.
2015-08-06 02:26:52 +02:00
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
ResultVal<Handle> KernelCore::CreateTimerCallbackHandle(const SharedPtr<Timer>& timer) {
return impl->timer_callback_handle_table.Create(timer);
}
CoreTiming::EventType* KernelCore::ThreadWakeupCallbackEventType() const {
return impl->thread_wakeup_event_type;
}
CoreTiming::EventType* KernelCore::TimerCallbackEventType() const {
return impl->timer_callback_event_type;
}
Kernel::HandleTable& KernelCore::ThreadWakeupCallbackHandleTable() {
return impl->thread_wakeup_callback_handle_table;
}
Kernel: Add more infrastructure to support different memory layouts This adds some structures necessary to support multiple memory regions in the future. It also adds support for different system memory types and the new linear heap mapping at 0x30000000.
2015-08-06 02:26:52 +02:00
kernel: Eliminate kernel global state As means to pave the way for getting rid of global state within core, This eliminates kernel global state by removing all globals. Instead this introduces a KernelCore class which acts as a kernel instance. This instance lives in the System class, which keeps its lifetime contained to the lifetime of the System class. This also forces the kernel types to actually interact with the main kernel instance itself instead of having transient kernel state placed all over several translation units, keeping everything together. It also has a nice consequence of making dependencies much more explicit. This also makes our initialization a tad bit more correct. Previously we were creating a kernel process before the actual kernel was initialized, which doesn't really make much sense. The KernelCore class itself follows the PImpl idiom, which allows keeping all the implementation details sealed away from everything else, which forces the use of the exposed API and allows us to avoid any unnecessary inclusions within the main kernel header.
2018-08-28 18:30:33 +02:00
const Kernel::HandleTable& KernelCore::ThreadWakeupCallbackHandleTable() const {
return impl->thread_wakeup_callback_handle_table;
- added __KernelLoadExec function - fixed some logging
2014-05-14 03:57:12 +02:00
}
kernel: refactored function naming to remove "__" prefix
2014-05-23 01:06:12 +02:00
kernel: Add ObjectAddressTable class.
2018-01-01 19:25:37 +01:00
} // namespace Kernel
Reference in New Issue Copy Permalink