2a1f59b301
svcCreateEvent operates by creating both a readable and writable event and then attempts to add both to the current process' handle table. If adding either of the events to the handle table fails, then the relevant error from the handle table is returned. If adding the readable event after the writable event to the table fails, then the writable event is removed from the handle table and the relevant error from the handle table is returned. Note that since we do not currently test resource limits, we don't check the resource limit table yet.
1926 lines
73 KiB
C++
1926 lines
73 KiB
C++
// Copyright 2018 yuzu emulator team
|
|
// Licensed under GPLv2 or any later version
|
|
// Refer to the license.txt file included.
|
|
|
|
#include <algorithm>
|
|
#include <cinttypes>
|
|
#include <iterator>
|
|
#include <mutex>
|
|
#include <vector>
|
|
|
|
#include "common/alignment.h"
|
|
#include "common/assert.h"
|
|
#include "common/logging/log.h"
|
|
#include "common/microprofile.h"
|
|
#include "common/string_util.h"
|
|
#include "core/arm/exclusive_monitor.h"
|
|
#include "core/core.h"
|
|
#include "core/core_cpu.h"
|
|
#include "core/core_timing.h"
|
|
#include "core/hle/kernel/address_arbiter.h"
|
|
#include "core/hle/kernel/client_port.h"
|
|
#include "core/hle/kernel/client_session.h"
|
|
#include "core/hle/kernel/handle_table.h"
|
|
#include "core/hle/kernel/kernel.h"
|
|
#include "core/hle/kernel/mutex.h"
|
|
#include "core/hle/kernel/process.h"
|
|
#include "core/hle/kernel/readable_event.h"
|
|
#include "core/hle/kernel/resource_limit.h"
|
|
#include "core/hle/kernel/scheduler.h"
|
|
#include "core/hle/kernel/shared_memory.h"
|
|
#include "core/hle/kernel/svc.h"
|
|
#include "core/hle/kernel/svc_wrap.h"
|
|
#include "core/hle/kernel/thread.h"
|
|
#include "core/hle/kernel/writable_event.h"
|
|
#include "core/hle/lock.h"
|
|
#include "core/hle/result.h"
|
|
#include "core/hle/service/service.h"
|
|
#include "core/settings.h"
|
|
|
|
namespace Kernel {
|
|
namespace {
|
|
|
|
// Checks if address + size is greater than the given address
|
|
// This can return false if the size causes an overflow of a 64-bit type
|
|
// or if the given size is zero.
|
|
constexpr bool IsValidAddressRange(VAddr address, u64 size) {
|
|
return address + size > address;
|
|
}
|
|
|
|
// Checks if a given address range lies within a larger address range.
|
|
constexpr bool IsInsideAddressRange(VAddr address, u64 size, VAddr address_range_begin,
|
|
VAddr address_range_end) {
|
|
const VAddr end_address = address + size - 1;
|
|
return address_range_begin <= address && end_address <= address_range_end - 1;
|
|
}
|
|
|
|
bool IsInsideAddressSpace(const VMManager& vm, VAddr address, u64 size) {
|
|
return IsInsideAddressRange(address, size, vm.GetAddressSpaceBaseAddress(),
|
|
vm.GetAddressSpaceEndAddress());
|
|
}
|
|
|
|
bool IsInsideNewMapRegion(const VMManager& vm, VAddr address, u64 size) {
|
|
return IsInsideAddressRange(address, size, vm.GetNewMapRegionBaseAddress(),
|
|
vm.GetNewMapRegionEndAddress());
|
|
}
|
|
|
|
// 8 GiB
|
|
constexpr u64 MAIN_MEMORY_SIZE = 0x200000000;
|
|
|
|
// Helper function that performs the common sanity checks for svcMapMemory
|
|
// and svcUnmapMemory. This is doable, as both functions perform their sanitizing
|
|
// in the same order.
|
|
ResultCode MapUnmapMemorySanityChecks(const VMManager& vm_manager, VAddr dst_addr, VAddr src_addr,
|
|
u64 size) {
|
|
if (!Common::Is4KBAligned(dst_addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Destination address is not aligned to 4KB, 0x{:016X}", dst_addr);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
if (!Common::Is4KBAligned(src_addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Source address is not aligned to 4KB, 0x{:016X}", src_addr);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (size == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Size is 0");
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!Common::Is4KBAligned(size)) {
|
|
LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, 0x{:016X}", size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!IsValidAddressRange(dst_addr, size)) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Destination is not a valid address range, addr=0x{:016X}, size=0x{:016X}",
|
|
dst_addr, size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
if (!IsValidAddressRange(src_addr, size)) {
|
|
LOG_ERROR(Kernel_SVC, "Source is not a valid address range, addr=0x{:016X}, size=0x{:016X}",
|
|
src_addr, size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
if (!IsInsideAddressSpace(vm_manager, src_addr, size)) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Source is not within the address space, addr=0x{:016X}, size=0x{:016X}",
|
|
src_addr, size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
if (!IsInsideNewMapRegion(vm_manager, dst_addr, size)) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Destination is not within the new map region, addr=0x{:016X}, size=0x{:016X}",
|
|
dst_addr, size);
|
|
return ERR_INVALID_MEMORY_RANGE;
|
|
}
|
|
|
|
const VAddr dst_end_address = dst_addr + size;
|
|
if (dst_end_address > vm_manager.GetHeapRegionBaseAddress() &&
|
|
vm_manager.GetHeapRegionEndAddress() > dst_addr) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Destination does not fit within the heap region, addr=0x{:016X}, "
|
|
"size=0x{:016X}, end_addr=0x{:016X}",
|
|
dst_addr, size, dst_end_address);
|
|
return ERR_INVALID_MEMORY_RANGE;
|
|
}
|
|
|
|
if (dst_end_address > vm_manager.GetMapRegionBaseAddress() &&
|
|
vm_manager.GetMapRegionEndAddress() > dst_addr) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Destination does not fit within the map region, addr=0x{:016X}, "
|
|
"size=0x{:016X}, end_addr=0x{:016X}",
|
|
dst_addr, size, dst_end_address);
|
|
return ERR_INVALID_MEMORY_RANGE;
|
|
}
|
|
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
enum class ResourceLimitValueType {
|
|
CurrentValue,
|
|
LimitValue,
|
|
};
|
|
|
|
ResultVal<s64> RetrieveResourceLimitValue(Handle resource_limit, u32 resource_type,
|
|
ResourceLimitValueType value_type) {
|
|
const auto type = static_cast<ResourceType>(resource_type);
|
|
if (!IsValidResourceType(type)) {
|
|
LOG_ERROR(Kernel_SVC, "Invalid resource limit type: '{}'", resource_type);
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
|
|
const auto& kernel = Core::System::GetInstance().Kernel();
|
|
const auto* const current_process = kernel.CurrentProcess();
|
|
ASSERT(current_process != nullptr);
|
|
|
|
const auto resource_limit_object =
|
|
current_process->GetHandleTable().Get<ResourceLimit>(resource_limit);
|
|
if (!resource_limit_object) {
|
|
LOG_ERROR(Kernel_SVC, "Handle to non-existent resource limit instance used. Handle={:08X}",
|
|
resource_limit);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
if (value_type == ResourceLimitValueType::CurrentValue) {
|
|
return MakeResult(resource_limit_object->GetCurrentResourceValue(type));
|
|
}
|
|
|
|
return MakeResult(resource_limit_object->GetMaxResourceValue(type));
|
|
}
|
|
} // Anonymous namespace
|
|
|
|
/// Set the process heap to a given Size. It can both extend and shrink the heap.
|
|
static ResultCode SetHeapSize(VAddr* heap_addr, u64 heap_size) {
|
|
LOG_TRACE(Kernel_SVC, "called, heap_size=0x{:X}", heap_size);
|
|
|
|
// Size must be a multiple of 0x200000 (2MB) and be equal to or less than 8GB.
|
|
if ((heap_size % 0x200000) != 0) {
|
|
LOG_ERROR(Kernel_SVC, "The heap size is not a multiple of 2MB, heap_size=0x{:016X}",
|
|
heap_size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (heap_size >= 0x200000000) {
|
|
LOG_ERROR(Kernel_SVC, "The heap size is not less than 8GB, heap_size=0x{:016X}", heap_size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
auto& process = *Core::CurrentProcess();
|
|
const VAddr heap_base = process.VMManager().GetHeapRegionBaseAddress();
|
|
CASCADE_RESULT(*heap_addr,
|
|
process.HeapAllocate(heap_base, heap_size, VMAPermission::ReadWrite));
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode SetMemoryPermission(VAddr addr, u64 size, u32 prot) {
|
|
LOG_TRACE(Kernel_SVC, "called, addr=0x{:X}, size=0x{:X}, prot=0x{:X}", addr, size, prot);
|
|
|
|
if (!Common::Is4KBAligned(addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Address is not aligned to 4KB, addr=0x{:016X}", addr);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
if (size == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Size is 0");
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!Common::Is4KBAligned(size)) {
|
|
LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, size=0x{:016X}", size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!IsValidAddressRange(addr, size)) {
|
|
LOG_ERROR(Kernel_SVC, "Region is not a valid address range, addr=0x{:016X}, size=0x{:016X}",
|
|
addr, size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
const auto permission = static_cast<MemoryPermission>(prot);
|
|
if (permission != MemoryPermission::None && permission != MemoryPermission::Read &&
|
|
permission != MemoryPermission::ReadWrite) {
|
|
LOG_ERROR(Kernel_SVC, "Invalid memory permission specified, Got memory permission=0x{:08X}",
|
|
static_cast<u32>(permission));
|
|
return ERR_INVALID_MEMORY_PERMISSIONS;
|
|
}
|
|
|
|
auto* const current_process = Core::CurrentProcess();
|
|
auto& vm_manager = current_process->VMManager();
|
|
|
|
if (!IsInsideAddressSpace(vm_manager, addr, size)) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Source is not within the address space, addr=0x{:016X}, size=0x{:016X}", addr,
|
|
size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
const VMManager::VMAHandle iter = vm_manager.FindVMA(addr);
|
|
if (iter == vm_manager.vma_map.end()) {
|
|
LOG_ERROR(Kernel_SVC, "Unable to find VMA for address=0x{:016X}", addr);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
LOG_WARNING(Kernel_SVC, "Uniformity check on protected memory is not implemented.");
|
|
// TODO: Performs a uniformity check to make sure only protected memory is changed (it doesn't
|
|
// make sense to allow changing permissions on kernel memory itself, etc).
|
|
|
|
const auto converted_permissions = SharedMemory::ConvertPermissions(permission);
|
|
|
|
return vm_manager.ReprotectRange(addr, size, converted_permissions);
|
|
}
|
|
|
|
static ResultCode SetMemoryAttribute(VAddr addr, u64 size, u32 state0, u32 state1) {
|
|
LOG_WARNING(Kernel_SVC,
|
|
"(STUBBED) called, addr=0x{:X}, size=0x{:X}, state0=0x{:X}, state1=0x{:X}", addr,
|
|
size, state0, state1);
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Maps a memory range into a different range.
|
|
static ResultCode MapMemory(VAddr dst_addr, VAddr src_addr, u64 size) {
|
|
LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr,
|
|
src_addr, size);
|
|
|
|
auto* const current_process = Core::CurrentProcess();
|
|
const auto& vm_manager = current_process->VMManager();
|
|
|
|
const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size);
|
|
if (result != RESULT_SUCCESS) {
|
|
return result;
|
|
}
|
|
|
|
return current_process->MirrorMemory(dst_addr, src_addr, size);
|
|
}
|
|
|
|
/// Unmaps a region that was previously mapped with svcMapMemory
|
|
static ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size) {
|
|
LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr,
|
|
src_addr, size);
|
|
|
|
auto* const current_process = Core::CurrentProcess();
|
|
const auto& vm_manager = current_process->VMManager();
|
|
|
|
const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size);
|
|
if (result != RESULT_SUCCESS) {
|
|
return result;
|
|
}
|
|
|
|
return current_process->UnmapMemory(dst_addr, src_addr, size);
|
|
}
|
|
|
|
/// Connect to an OS service given the port name, returns the handle to the port to out
|
|
static ResultCode ConnectToNamedPort(Handle* out_handle, VAddr port_name_address) {
|
|
if (!Memory::IsValidVirtualAddress(port_name_address)) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Port Name Address is not a valid virtual address, port_name_address=0x{:016X}",
|
|
port_name_address);
|
|
return ERR_NOT_FOUND;
|
|
}
|
|
|
|
static constexpr std::size_t PortNameMaxLength = 11;
|
|
// Read 1 char beyond the max allowed port name to detect names that are too long.
|
|
std::string port_name = Memory::ReadCString(port_name_address, PortNameMaxLength + 1);
|
|
if (port_name.size() > PortNameMaxLength) {
|
|
LOG_ERROR(Kernel_SVC, "Port name is too long, expected {} but got {}", PortNameMaxLength,
|
|
port_name.size());
|
|
return ERR_OUT_OF_RANGE;
|
|
}
|
|
|
|
LOG_TRACE(Kernel_SVC, "called port_name={}", port_name);
|
|
|
|
auto& kernel = Core::System::GetInstance().Kernel();
|
|
auto it = kernel.FindNamedPort(port_name);
|
|
if (!kernel.IsValidNamedPort(it)) {
|
|
LOG_WARNING(Kernel_SVC, "tried to connect to unknown port: {}", port_name);
|
|
return ERR_NOT_FOUND;
|
|
}
|
|
|
|
auto client_port = it->second;
|
|
|
|
SharedPtr<ClientSession> client_session;
|
|
CASCADE_RESULT(client_session, client_port->Connect());
|
|
|
|
// Return the client session
|
|
auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
CASCADE_RESULT(*out_handle, handle_table.Create(client_session));
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Makes a blocking IPC call to an OS service.
|
|
static ResultCode SendSyncRequest(Handle handle) {
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
SharedPtr<ClientSession> session = handle_table.Get<ClientSession>(handle);
|
|
if (!session) {
|
|
LOG_ERROR(Kernel_SVC, "called with invalid handle=0x{:08X}", handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
LOG_TRACE(Kernel_SVC, "called handle=0x{:08X}({})", handle, session->GetName());
|
|
|
|
Core::System::GetInstance().PrepareReschedule();
|
|
|
|
// TODO(Subv): svcSendSyncRequest should put the caller thread to sleep while the server
|
|
// responds and cause a reschedule.
|
|
return session->SendSyncRequest(GetCurrentThread());
|
|
}
|
|
|
|
/// Get the ID for the specified thread.
|
|
static ResultCode GetThreadId(u32* thread_id, Handle thread_handle) {
|
|
LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", thread_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
*thread_id = thread->GetThreadID();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Get the ID of the specified process
|
|
static ResultCode GetProcessId(u32* process_id, Handle process_handle) {
|
|
LOG_TRACE(Kernel_SVC, "called process=0x{:08X}", process_handle);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
|
|
if (!process) {
|
|
LOG_ERROR(Kernel_SVC, "Process handle does not exist, process_handle=0x{:08X}",
|
|
process_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
*process_id = process->GetProcessID();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Default thread wakeup callback for WaitSynchronization
|
|
static bool DefaultThreadWakeupCallback(ThreadWakeupReason reason, SharedPtr<Thread> thread,
|
|
SharedPtr<WaitObject> object, std::size_t index) {
|
|
ASSERT(thread->GetStatus() == ThreadStatus::WaitSynchAny);
|
|
|
|
if (reason == ThreadWakeupReason::Timeout) {
|
|
thread->SetWaitSynchronizationResult(RESULT_TIMEOUT);
|
|
return true;
|
|
}
|
|
|
|
ASSERT(reason == ThreadWakeupReason::Signal);
|
|
thread->SetWaitSynchronizationResult(RESULT_SUCCESS);
|
|
thread->SetWaitSynchronizationOutput(static_cast<u32>(index));
|
|
return true;
|
|
};
|
|
|
|
/// Wait for the given handles to synchronize, timeout after the specified nanoseconds
|
|
static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64 handle_count,
|
|
s64 nano_seconds) {
|
|
LOG_TRACE(Kernel_SVC, "called handles_address=0x{:X}, handle_count={}, nano_seconds={}",
|
|
handles_address, handle_count, nano_seconds);
|
|
|
|
if (!Memory::IsValidVirtualAddress(handles_address)) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Handle address is not a valid virtual address, handle_address=0x{:016X}",
|
|
handles_address);
|
|
return ERR_INVALID_POINTER;
|
|
}
|
|
|
|
static constexpr u64 MaxHandles = 0x40;
|
|
|
|
if (handle_count > MaxHandles) {
|
|
LOG_ERROR(Kernel_SVC, "Handle count specified is too large, expected {} but got {}",
|
|
MaxHandles, handle_count);
|
|
return ERR_OUT_OF_RANGE;
|
|
}
|
|
|
|
auto* const thread = GetCurrentThread();
|
|
|
|
using ObjectPtr = Thread::ThreadWaitObjects::value_type;
|
|
Thread::ThreadWaitObjects objects(handle_count);
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
|
|
for (u64 i = 0; i < handle_count; ++i) {
|
|
const Handle handle = Memory::Read32(handles_address + i * sizeof(Handle));
|
|
const auto object = handle_table.Get<WaitObject>(handle);
|
|
|
|
if (object == nullptr) {
|
|
LOG_ERROR(Kernel_SVC, "Object is a nullptr");
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
objects[i] = object;
|
|
}
|
|
|
|
// Find the first object that is acquirable in the provided list of objects
|
|
auto itr = std::find_if(objects.begin(), objects.end(), [thread](const ObjectPtr& object) {
|
|
return !object->ShouldWait(thread);
|
|
});
|
|
|
|
if (itr != objects.end()) {
|
|
// We found a ready object, acquire it and set the result value
|
|
WaitObject* object = itr->get();
|
|
object->Acquire(thread);
|
|
*index = static_cast<s32>(std::distance(objects.begin(), itr));
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
// No objects were ready to be acquired, prepare to suspend the thread.
|
|
|
|
// If a timeout value of 0 was provided, just return the Timeout error code instead of
|
|
// suspending the thread.
|
|
if (nano_seconds == 0) {
|
|
return RESULT_TIMEOUT;
|
|
}
|
|
|
|
for (auto& object : objects) {
|
|
object->AddWaitingThread(thread);
|
|
}
|
|
|
|
thread->SetWaitObjects(std::move(objects));
|
|
thread->SetStatus(ThreadStatus::WaitSynchAny);
|
|
|
|
// Create an event to wake the thread up after the specified nanosecond delay has passed
|
|
thread->WakeAfterDelay(nano_seconds);
|
|
thread->SetWakeupCallback(DefaultThreadWakeupCallback);
|
|
|
|
Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule();
|
|
|
|
return RESULT_TIMEOUT;
|
|
}
|
|
|
|
/// Resumes a thread waiting on WaitSynchronization
|
|
static ResultCode CancelSynchronization(Handle thread_handle) {
|
|
LOG_TRACE(Kernel_SVC, "called thread=0x{:X}", thread_handle);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
|
|
thread_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
ASSERT(thread->GetStatus() == ThreadStatus::WaitSynchAny);
|
|
thread->SetWaitSynchronizationResult(ERR_SYNCHRONIZATION_CANCELED);
|
|
thread->ResumeFromWait();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Attempts to locks a mutex, creating it if it does not already exist
|
|
static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr,
|
|
Handle requesting_thread_handle) {
|
|
LOG_TRACE(Kernel_SVC,
|
|
"called holding_thread_handle=0x{:08X}, mutex_addr=0x{:X}, "
|
|
"requesting_current_thread_handle=0x{:08X}",
|
|
holding_thread_handle, mutex_addr, requesting_thread_handle);
|
|
|
|
if (Memory::IsKernelVirtualAddress(mutex_addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Mutex Address is a kernel virtual address, mutex_addr={:016X}",
|
|
mutex_addr);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
if (!Common::IsWordAligned(mutex_addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Mutex Address is not word aligned, mutex_addr={:016X}", mutex_addr);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
return Mutex::TryAcquire(handle_table, mutex_addr, holding_thread_handle,
|
|
requesting_thread_handle);
|
|
}
|
|
|
|
/// Unlock a mutex
|
|
static ResultCode ArbitrateUnlock(VAddr mutex_addr) {
|
|
LOG_TRACE(Kernel_SVC, "called mutex_addr=0x{:X}", mutex_addr);
|
|
|
|
if (Memory::IsKernelVirtualAddress(mutex_addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Mutex Address is a kernel virtual address, mutex_addr={:016X}",
|
|
mutex_addr);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
if (!Common::IsWordAligned(mutex_addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Mutex Address is not word aligned, mutex_addr={:016X}", mutex_addr);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
return Mutex::Release(mutex_addr);
|
|
}
|
|
|
|
enum class BreakType : u32 {
|
|
Panic = 0,
|
|
AssertionFailed = 1,
|
|
PreNROLoad = 3,
|
|
PostNROLoad = 4,
|
|
PreNROUnload = 5,
|
|
PostNROUnload = 6,
|
|
};
|
|
|
|
struct BreakReason {
|
|
union {
|
|
u32 raw;
|
|
BitField<0, 30, BreakType> break_type;
|
|
BitField<31, 1, u32> signal_debugger;
|
|
};
|
|
};
|
|
|
|
/// Break program execution
|
|
static void Break(u32 reason, u64 info1, u64 info2) {
|
|
BreakReason break_reason{reason};
|
|
bool has_dumped_buffer{};
|
|
|
|
const auto handle_debug_buffer = [&](VAddr addr, u64 sz) {
|
|
if (sz == 0 || addr == 0 || has_dumped_buffer) {
|
|
return;
|
|
}
|
|
|
|
// This typically is an error code so we're going to assume this is the case
|
|
if (sz == sizeof(u32)) {
|
|
LOG_CRITICAL(Debug_Emulated, "debug_buffer_err_code={:X}", Memory::Read32(addr));
|
|
} else {
|
|
// We don't know what's in here so we'll hexdump it
|
|
std::vector<u8> debug_buffer(sz);
|
|
Memory::ReadBlock(addr, debug_buffer.data(), sz);
|
|
std::string hexdump;
|
|
for (std::size_t i = 0; i < debug_buffer.size(); i++) {
|
|
hexdump += fmt::format("{:02X} ", debug_buffer[i]);
|
|
if (i != 0 && i % 16 == 0) {
|
|
hexdump += '\n';
|
|
}
|
|
}
|
|
LOG_CRITICAL(Debug_Emulated, "debug_buffer=\n{}", hexdump);
|
|
}
|
|
has_dumped_buffer = true;
|
|
};
|
|
switch (break_reason.break_type) {
|
|
case BreakType::Panic:
|
|
LOG_CRITICAL(Debug_Emulated, "Signalling debugger, PANIC! info1=0x{:016X}, info2=0x{:016X}",
|
|
info1, info2);
|
|
handle_debug_buffer(info1, info2);
|
|
break;
|
|
case BreakType::AssertionFailed:
|
|
LOG_CRITICAL(Debug_Emulated,
|
|
"Signalling debugger, Assertion failed! info1=0x{:016X}, info2=0x{:016X}",
|
|
info1, info2);
|
|
handle_debug_buffer(info1, info2);
|
|
break;
|
|
case BreakType::PreNROLoad:
|
|
LOG_WARNING(
|
|
Debug_Emulated,
|
|
"Signalling debugger, Attempting to load an NRO at 0x{:016X} with size 0x{:016X}",
|
|
info1, info2);
|
|
break;
|
|
case BreakType::PostNROLoad:
|
|
LOG_WARNING(Debug_Emulated,
|
|
"Signalling debugger, Loaded an NRO at 0x{:016X} with size 0x{:016X}", info1,
|
|
info2);
|
|
break;
|
|
case BreakType::PreNROUnload:
|
|
LOG_WARNING(
|
|
Debug_Emulated,
|
|
"Signalling debugger, Attempting to unload an NRO at 0x{:016X} with size 0x{:016X}",
|
|
info1, info2);
|
|
break;
|
|
case BreakType::PostNROUnload:
|
|
LOG_WARNING(Debug_Emulated,
|
|
"Signalling debugger, Unloaded an NRO at 0x{:016X} with size 0x{:016X}", info1,
|
|
info2);
|
|
break;
|
|
default:
|
|
LOG_WARNING(
|
|
Debug_Emulated,
|
|
"Signalling debugger, Unknown break reason {}, info1=0x{:016X}, info2=0x{:016X}",
|
|
static_cast<u32>(break_reason.break_type.Value()), info1, info2);
|
|
handle_debug_buffer(info1, info2);
|
|
break;
|
|
}
|
|
|
|
if (!break_reason.signal_debugger) {
|
|
LOG_CRITICAL(
|
|
Debug_Emulated,
|
|
"Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
|
|
reason, info1, info2);
|
|
handle_debug_buffer(info1, info2);
|
|
ASSERT(false);
|
|
|
|
Core::CurrentProcess()->PrepareForTermination();
|
|
|
|
// Kill the current thread
|
|
GetCurrentThread()->Stop();
|
|
Core::System::GetInstance().PrepareReschedule();
|
|
}
|
|
}
|
|
|
|
/// Used to output a message on a debug hardware unit - does nothing on a retail unit
|
|
static void OutputDebugString(VAddr address, u64 len) {
|
|
if (len == 0) {
|
|
return;
|
|
}
|
|
|
|
std::string str(len, '\0');
|
|
Memory::ReadBlock(address, str.data(), str.size());
|
|
LOG_DEBUG(Debug_Emulated, "{}", str);
|
|
}
|
|
|
|
/// Gets system/memory information for the current process
|
|
static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) {
|
|
LOG_TRACE(Kernel_SVC, "called info_id=0x{:X}, info_sub_id=0x{:X}, handle=0x{:08X}", info_id,
|
|
info_sub_id, handle);
|
|
|
|
enum class GetInfoType : u64 {
|
|
// 1.0.0+
|
|
AllowedCpuIdBitmask = 0,
|
|
AllowedThreadPrioBitmask = 1,
|
|
MapRegionBaseAddr = 2,
|
|
MapRegionSize = 3,
|
|
HeapRegionBaseAddr = 4,
|
|
HeapRegionSize = 5,
|
|
TotalMemoryUsage = 6,
|
|
TotalHeapUsage = 7,
|
|
IsCurrentProcessBeingDebugged = 8,
|
|
RegisterResourceLimit = 9,
|
|
IdleTickCount = 10,
|
|
RandomEntropy = 11,
|
|
PerformanceCounter = 0xF0000002,
|
|
// 2.0.0+
|
|
ASLRRegionBaseAddr = 12,
|
|
ASLRRegionSize = 13,
|
|
NewMapRegionBaseAddr = 14,
|
|
NewMapRegionSize = 15,
|
|
// 3.0.0+
|
|
IsVirtualAddressMemoryEnabled = 16,
|
|
PersonalMmHeapUsage = 17,
|
|
TitleId = 18,
|
|
// 4.0.0+
|
|
PrivilegedProcessId = 19,
|
|
// 5.0.0+
|
|
UserExceptionContextAddr = 20,
|
|
ThreadTickCount = 0xF0000002,
|
|
};
|
|
|
|
const auto info_id_type = static_cast<GetInfoType>(info_id);
|
|
|
|
switch (info_id_type) {
|
|
case GetInfoType::AllowedCpuIdBitmask:
|
|
case GetInfoType::AllowedThreadPrioBitmask:
|
|
case GetInfoType::MapRegionBaseAddr:
|
|
case GetInfoType::MapRegionSize:
|
|
case GetInfoType::HeapRegionBaseAddr:
|
|
case GetInfoType::HeapRegionSize:
|
|
case GetInfoType::ASLRRegionBaseAddr:
|
|
case GetInfoType::ASLRRegionSize:
|
|
case GetInfoType::NewMapRegionBaseAddr:
|
|
case GetInfoType::NewMapRegionSize:
|
|
case GetInfoType::TotalMemoryUsage:
|
|
case GetInfoType::TotalHeapUsage:
|
|
case GetInfoType::IsVirtualAddressMemoryEnabled:
|
|
case GetInfoType::PersonalMmHeapUsage:
|
|
case GetInfoType::TitleId:
|
|
case GetInfoType::UserExceptionContextAddr: {
|
|
if (info_sub_id != 0) {
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
|
|
const auto& current_process_handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const auto process = current_process_handle_table.Get<Process>(static_cast<Handle>(handle));
|
|
if (!process) {
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
switch (info_id_type) {
|
|
case GetInfoType::AllowedCpuIdBitmask:
|
|
*result = process->GetAllowedProcessorMask();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::AllowedThreadPrioBitmask:
|
|
*result = process->GetAllowedThreadPriorityMask();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::MapRegionBaseAddr:
|
|
*result = process->VMManager().GetMapRegionBaseAddress();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::MapRegionSize:
|
|
*result = process->VMManager().GetMapRegionSize();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::HeapRegionBaseAddr:
|
|
*result = process->VMManager().GetHeapRegionBaseAddress();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::HeapRegionSize:
|
|
*result = process->VMManager().GetHeapRegionSize();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::ASLRRegionBaseAddr:
|
|
*result = process->VMManager().GetASLRRegionBaseAddress();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::ASLRRegionSize:
|
|
*result = process->VMManager().GetASLRRegionSize();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::NewMapRegionBaseAddr:
|
|
*result = process->VMManager().GetNewMapRegionBaseAddress();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::NewMapRegionSize:
|
|
*result = process->VMManager().GetNewMapRegionSize();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::TotalMemoryUsage:
|
|
*result = process->VMManager().GetTotalMemoryUsage();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::TotalHeapUsage:
|
|
*result = process->VMManager().GetTotalHeapUsage();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::IsVirtualAddressMemoryEnabled:
|
|
*result = process->IsVirtualMemoryEnabled();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::TitleId:
|
|
*result = process->GetTitleID();
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::UserExceptionContextAddr:
|
|
LOG_WARNING(Kernel_SVC,
|
|
"(STUBBED) Attempted to query user exception context address, returned 0");
|
|
*result = 0;
|
|
return RESULT_SUCCESS;
|
|
|
|
default:
|
|
break;
|
|
}
|
|
|
|
LOG_WARNING(Kernel_SVC, "(STUBBED) Unimplemented svcGetInfo id=0x{:016X}", info_id);
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
|
|
case GetInfoType::IsCurrentProcessBeingDebugged:
|
|
*result = 0;
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::RegisterResourceLimit: {
|
|
if (handle != 0) {
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
if (info_sub_id != 0) {
|
|
return ERR_INVALID_COMBINATION;
|
|
}
|
|
|
|
Process* const current_process = Core::CurrentProcess();
|
|
HandleTable& handle_table = current_process->GetHandleTable();
|
|
const auto resource_limit = current_process->GetResourceLimit();
|
|
if (!resource_limit) {
|
|
*result = KernelHandle::InvalidHandle;
|
|
// Yes, the kernel considers this a successful operation.
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
const auto table_result = handle_table.Create(resource_limit);
|
|
if (table_result.Failed()) {
|
|
return table_result.Code();
|
|
}
|
|
|
|
*result = *table_result;
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
case GetInfoType::RandomEntropy:
|
|
if (handle != 0) {
|
|
LOG_ERROR(Kernel_SVC, "Process Handle is non zero, expected 0 result but got {:016X}",
|
|
handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
if (info_sub_id >= Process::RANDOM_ENTROPY_SIZE) {
|
|
LOG_ERROR(Kernel_SVC, "Entropy size is out of range, expected {} but got {}",
|
|
Process::RANDOM_ENTROPY_SIZE, info_sub_id);
|
|
return ERR_INVALID_COMBINATION;
|
|
}
|
|
|
|
*result = Core::CurrentProcess()->GetRandomEntropy(info_sub_id);
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::PrivilegedProcessId:
|
|
LOG_WARNING(Kernel_SVC,
|
|
"(STUBBED) Attempted to query privileged process id bounds, returned 0");
|
|
*result = 0;
|
|
return RESULT_SUCCESS;
|
|
|
|
case GetInfoType::ThreadTickCount: {
|
|
constexpr u64 num_cpus = 4;
|
|
if (info_sub_id != 0xFFFFFFFFFFFFFFFF && info_sub_id >= num_cpus) {
|
|
LOG_ERROR(Kernel_SVC, "Core count is out of range, expected {} but got {}", num_cpus,
|
|
info_sub_id);
|
|
return ERR_INVALID_COMBINATION;
|
|
}
|
|
|
|
const auto thread =
|
|
Core::CurrentProcess()->GetHandleTable().Get<Thread>(static_cast<Handle>(handle));
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}",
|
|
static_cast<Handle>(handle));
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
const auto& system = Core::System::GetInstance();
|
|
const auto& scheduler = system.CurrentScheduler();
|
|
const auto* const current_thread = scheduler.GetCurrentThread();
|
|
const bool same_thread = current_thread == thread;
|
|
|
|
const u64 prev_ctx_ticks = scheduler.GetLastContextSwitchTicks();
|
|
u64 out_ticks = 0;
|
|
if (same_thread && info_sub_id == 0xFFFFFFFFFFFFFFFF) {
|
|
const u64 thread_ticks = current_thread->GetTotalCPUTimeTicks();
|
|
|
|
out_ticks = thread_ticks + (CoreTiming::GetTicks() - prev_ctx_ticks);
|
|
} else if (same_thread && info_sub_id == system.CurrentCoreIndex()) {
|
|
out_ticks = CoreTiming::GetTicks() - prev_ctx_ticks;
|
|
}
|
|
|
|
*result = out_ticks;
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
default:
|
|
LOG_WARNING(Kernel_SVC, "(STUBBED) Unimplemented svcGetInfo id=0x{:016X}", info_id);
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
}
|
|
|
|
/// Sets the thread activity
|
|
static ResultCode SetThreadActivity(Handle handle, u32 unknown) {
|
|
LOG_WARNING(Kernel_SVC, "(STUBBED) called, handle=0x{:08X}, unknown=0x{:08X}", handle, unknown);
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Gets the thread context
|
|
static ResultCode GetThreadContext(VAddr thread_context, Handle handle) {
|
|
LOG_DEBUG(Kernel_SVC, "called, context=0x{:08X}, thread=0x{:X}", thread_context, handle);
|
|
|
|
const auto* current_process = Core::CurrentProcess();
|
|
const SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
if (thread->GetOwnerProcess() != current_process) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"The current process does not own the current thread, thread_handle={:08X} "
|
|
"thread_pid={}, "
|
|
"current_process_pid={}",
|
|
handle, thread->GetOwnerProcess()->GetProcessID(),
|
|
current_process->GetProcessID());
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
if (thread == GetCurrentThread()) {
|
|
LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread");
|
|
return ERR_ALREADY_REGISTERED;
|
|
}
|
|
|
|
Core::ARM_Interface::ThreadContext ctx = thread->GetContext();
|
|
// Mask away mode bits, interrupt bits, IL bit, and other reserved bits.
|
|
ctx.pstate &= 0xFF0FFE20;
|
|
|
|
// If 64-bit, we can just write the context registers directly and we're good.
|
|
// However, if 32-bit, we have to ensure some registers are zeroed out.
|
|
if (!current_process->Is64BitProcess()) {
|
|
std::fill(ctx.cpu_registers.begin() + 15, ctx.cpu_registers.end(), 0);
|
|
std::fill(ctx.vector_registers.begin() + 16, ctx.vector_registers.end(), u128{});
|
|
}
|
|
|
|
Memory::WriteBlock(thread_context, &ctx, sizeof(ctx));
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Gets the priority for the specified thread
|
|
static ResultCode GetThreadPriority(u32* priority, Handle handle) {
|
|
LOG_TRACE(Kernel_SVC, "called");
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
*priority = thread->GetPriority();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Sets the priority for the specified thread
|
|
static ResultCode SetThreadPriority(Handle handle, u32 priority) {
|
|
LOG_TRACE(Kernel_SVC, "called");
|
|
|
|
if (priority > THREADPRIO_LOWEST) {
|
|
LOG_ERROR(
|
|
Kernel_SVC,
|
|
"An invalid priority was specified, expected {} but got {} for thread_handle={:08X}",
|
|
THREADPRIO_LOWEST, priority, handle);
|
|
return ERR_INVALID_THREAD_PRIORITY;
|
|
}
|
|
|
|
const auto* const current_process = Core::CurrentProcess();
|
|
|
|
SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
thread->SetPriority(priority);
|
|
|
|
Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Get which CPU core is executing the current thread
|
|
static u32 GetCurrentProcessorNumber() {
|
|
LOG_TRACE(Kernel_SVC, "called");
|
|
return GetCurrentThread()->GetProcessorID();
|
|
}
|
|
|
|
static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 size,
|
|
u32 permissions) {
|
|
LOG_TRACE(Kernel_SVC,
|
|
"called, shared_memory_handle=0x{:X}, addr=0x{:X}, size=0x{:X}, permissions=0x{:08X}",
|
|
shared_memory_handle, addr, size, permissions);
|
|
|
|
if (!Common::Is4KBAligned(addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Address is not aligned to 4KB, addr=0x{:016X}", addr);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
if (size == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Size is 0");
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!Common::Is4KBAligned(size)) {
|
|
LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, size=0x{:016X}", size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!IsValidAddressRange(addr, size)) {
|
|
LOG_ERROR(Kernel_SVC, "Region is not a valid address range, addr=0x{:016X}, size=0x{:016X}",
|
|
addr, size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
const auto permissions_type = static_cast<MemoryPermission>(permissions);
|
|
if (permissions_type != MemoryPermission::Read &&
|
|
permissions_type != MemoryPermission::ReadWrite) {
|
|
LOG_ERROR(Kernel_SVC, "Expected Read or ReadWrite permission but got permissions=0x{:08X}",
|
|
permissions);
|
|
return ERR_INVALID_MEMORY_PERMISSIONS;
|
|
}
|
|
|
|
auto* const current_process = Core::CurrentProcess();
|
|
auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle);
|
|
if (!shared_memory) {
|
|
LOG_ERROR(Kernel_SVC, "Shared memory does not exist, shared_memory_handle=0x{:08X}",
|
|
shared_memory_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
const auto& vm_manager = current_process->VMManager();
|
|
if (!vm_manager.IsWithinASLRRegion(addr, size)) {
|
|
LOG_ERROR(Kernel_SVC, "Region is not within the ASLR region. addr=0x{:016X}, size={:016X}",
|
|
addr, size);
|
|
return ERR_INVALID_MEMORY_RANGE;
|
|
}
|
|
|
|
return shared_memory->Map(*current_process, addr, permissions_type, MemoryPermission::DontCare);
|
|
}
|
|
|
|
static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 size) {
|
|
LOG_WARNING(Kernel_SVC, "called, shared_memory_handle=0x{:08X}, addr=0x{:X}, size=0x{:X}",
|
|
shared_memory_handle, addr, size);
|
|
|
|
if (!Common::Is4KBAligned(addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Address is not aligned to 4KB, addr=0x{:016X}", addr);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
if (size == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Size is 0");
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!Common::Is4KBAligned(size)) {
|
|
LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, size=0x{:016X}", size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (!IsValidAddressRange(addr, size)) {
|
|
LOG_ERROR(Kernel_SVC, "Region is not a valid address range, addr=0x{:016X}, size=0x{:016X}",
|
|
addr, size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
auto* const current_process = Core::CurrentProcess();
|
|
auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle);
|
|
if (!shared_memory) {
|
|
LOG_ERROR(Kernel_SVC, "Shared memory does not exist, shared_memory_handle=0x{:08X}",
|
|
shared_memory_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
const auto& vm_manager = current_process->VMManager();
|
|
if (!vm_manager.IsWithinASLRRegion(addr, size)) {
|
|
LOG_ERROR(Kernel_SVC, "Region is not within the ASLR region. addr=0x{:016X}, size={:016X}",
|
|
addr, size);
|
|
return ERR_INVALID_MEMORY_RANGE;
|
|
}
|
|
|
|
return shared_memory->Unmap(*current_process, addr);
|
|
}
|
|
|
|
/// Query process memory
|
|
static ResultCode QueryProcessMemory(MemoryInfo* memory_info, PageInfo* /*page_info*/,
|
|
Handle process_handle, u64 addr) {
|
|
LOG_TRACE(Kernel_SVC, "called process=0x{:08X} addr={:X}", process_handle, addr);
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
|
|
if (!process) {
|
|
LOG_ERROR(Kernel_SVC, "Process handle does not exist, process_handle=0x{:08X}",
|
|
process_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
auto vma = process->VMManager().FindVMA(addr);
|
|
memory_info->attributes = 0;
|
|
if (vma == process->VMManager().vma_map.end()) {
|
|
memory_info->base_address = 0;
|
|
memory_info->permission = static_cast<u32>(VMAPermission::None);
|
|
memory_info->size = 0;
|
|
memory_info->type = static_cast<u32>(MemoryState::Unmapped);
|
|
} else {
|
|
memory_info->base_address = vma->second.base;
|
|
memory_info->permission = static_cast<u32>(vma->second.permissions);
|
|
memory_info->size = vma->second.size;
|
|
memory_info->type = static_cast<u32>(vma->second.meminfo_state);
|
|
}
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Query memory
|
|
static ResultCode QueryMemory(MemoryInfo* memory_info, PageInfo* page_info, VAddr addr) {
|
|
LOG_TRACE(Kernel_SVC, "called, addr={:X}", addr);
|
|
return QueryProcessMemory(memory_info, page_info, CurrentProcess, addr);
|
|
}
|
|
|
|
/// Exits the current process
|
|
static void ExitProcess() {
|
|
auto* current_process = Core::CurrentProcess();
|
|
|
|
LOG_INFO(Kernel_SVC, "Process {} exiting", current_process->GetProcessID());
|
|
ASSERT_MSG(current_process->GetStatus() == ProcessStatus::Running,
|
|
"Process has already exited");
|
|
|
|
current_process->PrepareForTermination();
|
|
|
|
// Kill the current thread
|
|
GetCurrentThread()->Stop();
|
|
|
|
Core::System::GetInstance().PrepareReschedule();
|
|
}
|
|
|
|
/// Creates a new thread
|
|
static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, VAddr stack_top,
|
|
u32 priority, s32 processor_id) {
|
|
LOG_TRACE(Kernel_SVC,
|
|
"called entrypoint=0x{:08X}, arg=0x{:08X}, stacktop=0x{:08X}, "
|
|
"threadpriority=0x{:08X}, processorid=0x{:08X} : created handle=0x{:08X}",
|
|
entry_point, arg, stack_top, priority, processor_id, *out_handle);
|
|
|
|
if (priority > THREADPRIO_LOWEST) {
|
|
LOG_ERROR(Kernel_SVC, "An invalid priority was specified, expected {} but got {}",
|
|
THREADPRIO_LOWEST, priority);
|
|
return ERR_INVALID_THREAD_PRIORITY;
|
|
}
|
|
|
|
auto* const current_process = Core::CurrentProcess();
|
|
|
|
if (processor_id == THREADPROCESSORID_DEFAULT) {
|
|
// Set the target CPU to the one specified in the process' exheader.
|
|
processor_id = current_process->GetDefaultProcessorID();
|
|
ASSERT(processor_id != THREADPROCESSORID_DEFAULT);
|
|
}
|
|
|
|
switch (processor_id) {
|
|
case THREADPROCESSORID_0:
|
|
case THREADPROCESSORID_1:
|
|
case THREADPROCESSORID_2:
|
|
case THREADPROCESSORID_3:
|
|
break;
|
|
default:
|
|
LOG_ERROR(Kernel_SVC, "Invalid thread processor ID: {}", processor_id);
|
|
return ERR_INVALID_PROCESSOR_ID;
|
|
}
|
|
|
|
const std::string name = fmt::format("thread-{:X}", entry_point);
|
|
auto& kernel = Core::System::GetInstance().Kernel();
|
|
CASCADE_RESULT(SharedPtr<Thread> thread,
|
|
Thread::Create(kernel, name, entry_point, priority, arg, processor_id, stack_top,
|
|
*current_process));
|
|
|
|
const auto new_guest_handle = current_process->GetHandleTable().Create(thread);
|
|
if (new_guest_handle.Failed()) {
|
|
LOG_ERROR(Kernel_SVC, "Failed to create handle with error=0x{:X}",
|
|
new_guest_handle.Code().raw);
|
|
return new_guest_handle.Code();
|
|
}
|
|
thread->SetGuestHandle(*new_guest_handle);
|
|
*out_handle = *new_guest_handle;
|
|
|
|
Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule();
|
|
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Starts the thread for the provided handle
|
|
static ResultCode StartThread(Handle thread_handle) {
|
|
LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
|
|
thread_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
ASSERT(thread->GetStatus() == ThreadStatus::Dormant);
|
|
|
|
thread->ResumeFromWait();
|
|
Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule();
|
|
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Called when a thread exits
|
|
static void ExitThread() {
|
|
LOG_TRACE(Kernel_SVC, "called, pc=0x{:08X}", Core::CurrentArmInterface().GetPC());
|
|
|
|
ExitCurrentThread();
|
|
Core::System::GetInstance().PrepareReschedule();
|
|
}
|
|
|
|
/// Sleep the current thread
|
|
static void SleepThread(s64 nanoseconds) {
|
|
LOG_TRACE(Kernel_SVC, "called nanoseconds={}", nanoseconds);
|
|
|
|
// Don't attempt to yield execution if there are no available threads to run,
|
|
// this way we avoid a useless reschedule to the idle thread.
|
|
if (nanoseconds == 0 && !Core::System::GetInstance().CurrentScheduler().HaveReadyThreads())
|
|
return;
|
|
|
|
// Sleep current thread and check for next thread to schedule
|
|
WaitCurrentThread_Sleep();
|
|
|
|
// Create an event to wake the thread up after the specified nanosecond delay has passed
|
|
GetCurrentThread()->WakeAfterDelay(nanoseconds);
|
|
|
|
Core::System::GetInstance().PrepareReschedule();
|
|
}
|
|
|
|
/// Wait process wide key atomic
|
|
static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_variable_addr,
|
|
Handle thread_handle, s64 nano_seconds) {
|
|
LOG_TRACE(
|
|
Kernel_SVC,
|
|
"called mutex_addr={:X}, condition_variable_addr={:X}, thread_handle=0x{:08X}, timeout={}",
|
|
mutex_addr, condition_variable_addr, thread_handle, nano_seconds);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
|
ASSERT(thread);
|
|
|
|
CASCADE_CODE(Mutex::Release(mutex_addr));
|
|
|
|
SharedPtr<Thread> current_thread = GetCurrentThread();
|
|
current_thread->SetCondVarWaitAddress(condition_variable_addr);
|
|
current_thread->SetMutexWaitAddress(mutex_addr);
|
|
current_thread->SetWaitHandle(thread_handle);
|
|
current_thread->SetStatus(ThreadStatus::WaitMutex);
|
|
current_thread->InvalidateWakeupCallback();
|
|
|
|
current_thread->WakeAfterDelay(nano_seconds);
|
|
|
|
// Note: Deliberately don't attempt to inherit the lock owner's priority.
|
|
|
|
Core::System::GetInstance().CpuCore(current_thread->GetProcessorID()).PrepareReschedule();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Signal process wide key
|
|
static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target) {
|
|
LOG_TRACE(Kernel_SVC, "called, condition_variable_addr=0x{:X}, target=0x{:08X}",
|
|
condition_variable_addr, target);
|
|
|
|
const auto RetrieveWaitingThreads = [](std::size_t core_index,
|
|
std::vector<SharedPtr<Thread>>& waiting_threads,
|
|
VAddr condvar_addr) {
|
|
const auto& scheduler = Core::System::GetInstance().Scheduler(core_index);
|
|
const auto& thread_list = scheduler.GetThreadList();
|
|
|
|
for (const auto& thread : thread_list) {
|
|
if (thread->GetCondVarWaitAddress() == condvar_addr)
|
|
waiting_threads.push_back(thread);
|
|
}
|
|
};
|
|
|
|
// Retrieve a list of all threads that are waiting for this condition variable.
|
|
std::vector<SharedPtr<Thread>> waiting_threads;
|
|
RetrieveWaitingThreads(0, waiting_threads, condition_variable_addr);
|
|
RetrieveWaitingThreads(1, waiting_threads, condition_variable_addr);
|
|
RetrieveWaitingThreads(2, waiting_threads, condition_variable_addr);
|
|
RetrieveWaitingThreads(3, waiting_threads, condition_variable_addr);
|
|
// Sort them by priority, such that the highest priority ones come first.
|
|
std::sort(waiting_threads.begin(), waiting_threads.end(),
|
|
[](const SharedPtr<Thread>& lhs, const SharedPtr<Thread>& rhs) {
|
|
return lhs->GetPriority() < rhs->GetPriority();
|
|
});
|
|
|
|
// Only process up to 'target' threads, unless 'target' is -1, in which case process
|
|
// them all.
|
|
std::size_t last = waiting_threads.size();
|
|
if (target != -1)
|
|
last = target;
|
|
|
|
// If there are no threads waiting on this condition variable, just exit
|
|
if (last > waiting_threads.size())
|
|
return RESULT_SUCCESS;
|
|
|
|
for (std::size_t index = 0; index < last; ++index) {
|
|
auto& thread = waiting_threads[index];
|
|
|
|
ASSERT(thread->GetCondVarWaitAddress() == condition_variable_addr);
|
|
|
|
std::size_t current_core = Core::System::GetInstance().CurrentCoreIndex();
|
|
|
|
auto& monitor = Core::System::GetInstance().Monitor();
|
|
|
|
// Atomically read the value of the mutex.
|
|
u32 mutex_val = 0;
|
|
do {
|
|
monitor.SetExclusive(current_core, thread->GetMutexWaitAddress());
|
|
|
|
// If the mutex is not yet acquired, acquire it.
|
|
mutex_val = Memory::Read32(thread->GetMutexWaitAddress());
|
|
|
|
if (mutex_val != 0) {
|
|
monitor.ClearExclusive();
|
|
break;
|
|
}
|
|
} while (!monitor.ExclusiveWrite32(current_core, thread->GetMutexWaitAddress(),
|
|
thread->GetWaitHandle()));
|
|
|
|
if (mutex_val == 0) {
|
|
// We were able to acquire the mutex, resume this thread.
|
|
ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex);
|
|
thread->ResumeFromWait();
|
|
|
|
auto* const lock_owner = thread->GetLockOwner();
|
|
if (lock_owner != nullptr) {
|
|
lock_owner->RemoveMutexWaiter(thread);
|
|
}
|
|
|
|
thread->SetLockOwner(nullptr);
|
|
thread->SetMutexWaitAddress(0);
|
|
thread->SetCondVarWaitAddress(0);
|
|
thread->SetWaitHandle(0);
|
|
} else {
|
|
// Atomically signal that the mutex now has a waiting thread.
|
|
do {
|
|
monitor.SetExclusive(current_core, thread->GetMutexWaitAddress());
|
|
|
|
// Ensure that the mutex value is still what we expect.
|
|
u32 value = Memory::Read32(thread->GetMutexWaitAddress());
|
|
// TODO(Subv): When this happens, the kernel just clears the exclusive state and
|
|
// retries the initial read for this thread.
|
|
ASSERT_MSG(mutex_val == value, "Unhandled synchronization primitive case");
|
|
} while (!monitor.ExclusiveWrite32(current_core, thread->GetMutexWaitAddress(),
|
|
mutex_val | Mutex::MutexHasWaitersFlag));
|
|
|
|
// The mutex is already owned by some other thread, make this thread wait on it.
|
|
const Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask);
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
auto owner = handle_table.Get<Thread>(owner_handle);
|
|
ASSERT(owner);
|
|
ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex);
|
|
thread->InvalidateWakeupCallback();
|
|
|
|
owner->AddMutexWaiter(thread);
|
|
|
|
Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule();
|
|
}
|
|
}
|
|
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
// Wait for an address (via Address Arbiter)
|
|
static ResultCode WaitForAddress(VAddr address, u32 type, s32 value, s64 timeout) {
|
|
LOG_WARNING(Kernel_SVC, "called, address=0x{:X}, type=0x{:X}, value=0x{:X}, timeout={}",
|
|
address, type, value, timeout);
|
|
// If the passed address is a kernel virtual address, return invalid memory state.
|
|
if (Memory::IsKernelVirtualAddress(address)) {
|
|
LOG_ERROR(Kernel_SVC, "Address is a kernel virtual address, address={:016X}", address);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
// If the address is not properly aligned to 4 bytes, return invalid address.
|
|
if (!Common::IsWordAligned(address)) {
|
|
LOG_ERROR(Kernel_SVC, "Address is not word aligned, address={:016X}", address);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
switch (static_cast<AddressArbiter::ArbitrationType>(type)) {
|
|
case AddressArbiter::ArbitrationType::WaitIfLessThan:
|
|
return AddressArbiter::WaitForAddressIfLessThan(address, value, timeout, false);
|
|
case AddressArbiter::ArbitrationType::DecrementAndWaitIfLessThan:
|
|
return AddressArbiter::WaitForAddressIfLessThan(address, value, timeout, true);
|
|
case AddressArbiter::ArbitrationType::WaitIfEqual:
|
|
return AddressArbiter::WaitForAddressIfEqual(address, value, timeout);
|
|
default:
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Invalid arbitration type, expected WaitIfLessThan, DecrementAndWaitIfLessThan "
|
|
"or WaitIfEqual but got {}",
|
|
type);
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
}
|
|
|
|
// Signals to an address (via Address Arbiter)
|
|
static ResultCode SignalToAddress(VAddr address, u32 type, s32 value, s32 num_to_wake) {
|
|
LOG_WARNING(Kernel_SVC, "called, address=0x{:X}, type=0x{:X}, value=0x{:X}, num_to_wake=0x{:X}",
|
|
address, type, value, num_to_wake);
|
|
// If the passed address is a kernel virtual address, return invalid memory state.
|
|
if (Memory::IsKernelVirtualAddress(address)) {
|
|
LOG_ERROR(Kernel_SVC, "Address is a kernel virtual address, address={:016X}", address);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
// If the address is not properly aligned to 4 bytes, return invalid address.
|
|
if (!Common::IsWordAligned(address)) {
|
|
LOG_ERROR(Kernel_SVC, "Address is not word aligned, address={:016X}", address);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
switch (static_cast<AddressArbiter::SignalType>(type)) {
|
|
case AddressArbiter::SignalType::Signal:
|
|
return AddressArbiter::SignalToAddress(address, num_to_wake);
|
|
case AddressArbiter::SignalType::IncrementAndSignalIfEqual:
|
|
return AddressArbiter::IncrementAndSignalToAddressIfEqual(address, value, num_to_wake);
|
|
case AddressArbiter::SignalType::ModifyByWaitingCountAndSignalIfEqual:
|
|
return AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(address, value,
|
|
num_to_wake);
|
|
default:
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Invalid signal type, expected Signal, IncrementAndSignalIfEqual "
|
|
"or ModifyByWaitingCountAndSignalIfEqual but got {}",
|
|
type);
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
}
|
|
|
|
/// This returns the total CPU ticks elapsed since the CPU was powered-on
|
|
static u64 GetSystemTick() {
|
|
LOG_TRACE(Kernel_SVC, "called");
|
|
|
|
const u64 result{CoreTiming::GetTicks()};
|
|
|
|
// Advance time to defeat dumb games that busy-wait for the frame to end.
|
|
CoreTiming::AddTicks(400);
|
|
|
|
return result;
|
|
}
|
|
|
|
/// Close a handle
|
|
static ResultCode CloseHandle(Handle handle) {
|
|
LOG_TRACE(Kernel_SVC, "Closing handle 0x{:08X}", handle);
|
|
|
|
auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
return handle_table.Close(handle);
|
|
}
|
|
|
|
/// Reset an event
|
|
static ResultCode ResetSignal(Handle handle) {
|
|
LOG_DEBUG(Kernel_SVC, "called handle 0x{:08X}", handle);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
auto event = handle_table.Get<ReadableEvent>(handle);
|
|
|
|
ASSERT(event != nullptr);
|
|
|
|
event->Clear();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
/// Creates a TransferMemory object
|
|
static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32 permissions) {
|
|
LOG_DEBUG(Kernel_SVC, "called addr=0x{:X}, size=0x{:X}, perms=0x{:08X}", addr, size,
|
|
permissions);
|
|
|
|
if (!Common::Is4KBAligned(addr)) {
|
|
LOG_ERROR(Kernel_SVC, "Address ({:016X}) is not page aligned!", addr);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
if (!Common::Is4KBAligned(size) || size == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Size ({:016X}) is not page aligned or equal to zero!", size);
|
|
return ERR_INVALID_ADDRESS;
|
|
}
|
|
|
|
if (!IsValidAddressRange(addr, size)) {
|
|
LOG_ERROR(Kernel_SVC, "Address and size cause overflow! (address={:016X}, size={:016X})",
|
|
addr, size);
|
|
return ERR_INVALID_ADDRESS_STATE;
|
|
}
|
|
|
|
const auto perms = static_cast<MemoryPermission>(permissions);
|
|
if (perms != MemoryPermission::None && perms != MemoryPermission::Read &&
|
|
perms != MemoryPermission::ReadWrite) {
|
|
LOG_ERROR(Kernel_SVC, "Invalid memory permissions for transfer memory! (perms={:08X})",
|
|
permissions);
|
|
return ERR_INVALID_MEMORY_PERMISSIONS;
|
|
}
|
|
|
|
auto& kernel = Core::System::GetInstance().Kernel();
|
|
auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const auto shared_mem_handle = SharedMemory::Create(
|
|
kernel, handle_table.Get<Process>(CurrentProcess), size, perms, perms, addr);
|
|
|
|
CASCADE_RESULT(*handle, handle_table.Create(shared_mem_handle));
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode GetThreadCoreMask(Handle thread_handle, u32* core, u64* mask) {
|
|
LOG_TRACE(Kernel_SVC, "called, handle=0x{:08X}", thread_handle);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
|
|
thread_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
*core = thread->GetIdealCore();
|
|
*mask = thread->GetAffinityMask();
|
|
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode SetThreadCoreMask(Handle thread_handle, u32 core, u64 mask) {
|
|
LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, mask=0x{:016X}, core=0x{:X}", thread_handle,
|
|
mask, core);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
|
|
if (!thread) {
|
|
LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}",
|
|
thread_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
if (core == static_cast<u32>(THREADPROCESSORID_DEFAULT)) {
|
|
const u8 default_processor_id = thread->GetOwnerProcess()->GetDefaultProcessorID();
|
|
|
|
ASSERT(default_processor_id != static_cast<u8>(THREADPROCESSORID_DEFAULT));
|
|
|
|
// Set the target CPU to the one specified in the process' exheader.
|
|
core = default_processor_id;
|
|
mask = 1ULL << core;
|
|
}
|
|
|
|
if (mask == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Mask is 0");
|
|
return ERR_INVALID_COMBINATION;
|
|
}
|
|
|
|
/// This value is used to only change the affinity mask without changing the current ideal core.
|
|
static constexpr u32 OnlyChangeMask = static_cast<u32>(-3);
|
|
|
|
if (core == OnlyChangeMask) {
|
|
core = thread->GetIdealCore();
|
|
} else if (core >= Core::NUM_CPU_CORES && core != static_cast<u32>(-1)) {
|
|
LOG_ERROR(Kernel_SVC, "Invalid core specified, got {}", core);
|
|
return ERR_INVALID_PROCESSOR_ID;
|
|
}
|
|
|
|
// Error out if the input core isn't enabled in the input mask.
|
|
if (core < Core::NUM_CPU_CORES && (mask & (1ull << core)) == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Core is not enabled for the current mask, core={}, mask={:016X}",
|
|
core, mask);
|
|
return ERR_INVALID_COMBINATION;
|
|
}
|
|
|
|
thread->ChangeCore(core, mask);
|
|
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permissions,
|
|
u32 remote_permissions) {
|
|
LOG_TRACE(Kernel_SVC, "called, size=0x{:X}, localPerms=0x{:08X}, remotePerms=0x{:08X}", size,
|
|
local_permissions, remote_permissions);
|
|
if (size == 0) {
|
|
LOG_ERROR(Kernel_SVC, "Size is 0");
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
if (!Common::Is4KBAligned(size)) {
|
|
LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, 0x{:016X}", size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
if (size >= MAIN_MEMORY_SIZE) {
|
|
LOG_ERROR(Kernel_SVC, "Size is not less than 8GB, 0x{:016X}", size);
|
|
return ERR_INVALID_SIZE;
|
|
}
|
|
|
|
const auto local_perms = static_cast<MemoryPermission>(local_permissions);
|
|
if (local_perms != MemoryPermission::Read && local_perms != MemoryPermission::ReadWrite) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Invalid local memory permissions, expected Read or ReadWrite but got "
|
|
"local_permissions={}",
|
|
static_cast<u32>(local_permissions));
|
|
return ERR_INVALID_MEMORY_PERMISSIONS;
|
|
}
|
|
|
|
const auto remote_perms = static_cast<MemoryPermission>(remote_permissions);
|
|
if (remote_perms != MemoryPermission::Read && remote_perms != MemoryPermission::ReadWrite &&
|
|
remote_perms != MemoryPermission::DontCare) {
|
|
LOG_ERROR(Kernel_SVC,
|
|
"Invalid remote memory permissions, expected Read, ReadWrite or DontCare but got "
|
|
"remote_permissions={}",
|
|
static_cast<u32>(remote_permissions));
|
|
return ERR_INVALID_MEMORY_PERMISSIONS;
|
|
}
|
|
|
|
auto& kernel = Core::System::GetInstance().Kernel();
|
|
auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
auto shared_mem_handle =
|
|
SharedMemory::Create(kernel, handle_table.Get<Process>(KernelHandle::CurrentProcess), size,
|
|
local_perms, remote_perms);
|
|
|
|
CASCADE_RESULT(*handle, handle_table.Create(shared_mem_handle));
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode CreateEvent(Handle* write_handle, Handle* read_handle) {
|
|
LOG_DEBUG(Kernel_SVC, "called");
|
|
|
|
auto& kernel = Core::System::GetInstance().Kernel();
|
|
const auto [readable_event, writable_event] =
|
|
WritableEvent::CreateEventPair(kernel, ResetType::Sticky, "CreateEvent");
|
|
|
|
HandleTable& handle_table = kernel.CurrentProcess()->GetHandleTable();
|
|
|
|
const auto write_create_result = handle_table.Create(writable_event);
|
|
if (write_create_result.Failed()) {
|
|
return write_create_result.Code();
|
|
}
|
|
*write_handle = *write_create_result;
|
|
|
|
const auto read_create_result = handle_table.Create(readable_event);
|
|
if (read_create_result.Failed()) {
|
|
handle_table.Close(*write_create_result);
|
|
return read_create_result.Code();
|
|
}
|
|
*read_handle = *read_create_result;
|
|
|
|
LOG_DEBUG(Kernel_SVC,
|
|
"successful. Writable event handle=0x{:08X}, Readable event handle=0x{:08X}",
|
|
*write_create_result, *read_create_result);
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode ClearEvent(Handle handle) {
|
|
LOG_TRACE(Kernel_SVC, "called, event=0x{:08X}", handle);
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
|
|
auto writable_event = handle_table.Get<WritableEvent>(handle);
|
|
if (writable_event) {
|
|
writable_event->Clear();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
auto readable_event = handle_table.Get<ReadableEvent>(handle);
|
|
if (readable_event) {
|
|
readable_event->Clear();
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
LOG_ERROR(Kernel_SVC, "Event handle does not exist, handle=0x{:08X}", handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
static ResultCode GetProcessInfo(u64* out, Handle process_handle, u32 type) {
|
|
LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, type=0x{:X}", process_handle, type);
|
|
|
|
// This function currently only allows retrieving a process' status.
|
|
enum class InfoType {
|
|
Status,
|
|
};
|
|
|
|
const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
|
|
const auto process = handle_table.Get<Process>(process_handle);
|
|
if (!process) {
|
|
LOG_ERROR(Kernel_SVC, "Process handle does not exist, process_handle=0x{:08X}",
|
|
process_handle);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
const auto info_type = static_cast<InfoType>(type);
|
|
if (info_type != InfoType::Status) {
|
|
LOG_ERROR(Kernel_SVC, "Expected info_type to be Status but got {} instead", type);
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
|
|
*out = static_cast<u64>(process->GetStatus());
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode CreateResourceLimit(Handle* out_handle) {
|
|
LOG_DEBUG(Kernel_SVC, "called");
|
|
|
|
auto& kernel = Core::System::GetInstance().Kernel();
|
|
auto resource_limit = ResourceLimit::Create(kernel);
|
|
|
|
auto* const current_process = kernel.CurrentProcess();
|
|
ASSERT(current_process != nullptr);
|
|
|
|
const auto handle = current_process->GetHandleTable().Create(std::move(resource_limit));
|
|
if (handle.Failed()) {
|
|
return handle.Code();
|
|
}
|
|
|
|
*out_handle = *handle;
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode GetResourceLimitLimitValue(u64* out_value, Handle resource_limit,
|
|
u32 resource_type) {
|
|
LOG_DEBUG(Kernel_SVC, "called. Handle={:08X}, Resource type={}", resource_limit, resource_type);
|
|
|
|
const auto limit_value = RetrieveResourceLimitValue(resource_limit, resource_type,
|
|
ResourceLimitValueType::LimitValue);
|
|
if (limit_value.Failed()) {
|
|
return limit_value.Code();
|
|
}
|
|
|
|
*out_value = static_cast<u64>(*limit_value);
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode GetResourceLimitCurrentValue(u64* out_value, Handle resource_limit,
|
|
u32 resource_type) {
|
|
LOG_DEBUG(Kernel_SVC, "called. Handle={:08X}, Resource type={}", resource_limit, resource_type);
|
|
|
|
const auto current_value = RetrieveResourceLimitValue(resource_limit, resource_type,
|
|
ResourceLimitValueType::CurrentValue);
|
|
if (current_value.Failed()) {
|
|
return current_value.Code();
|
|
}
|
|
|
|
*out_value = static_cast<u64>(*current_value);
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
static ResultCode SetResourceLimitLimitValue(Handle resource_limit, u32 resource_type, u64 value) {
|
|
LOG_DEBUG(Kernel_SVC, "called. Handle={:08X}, Resource type={}, Value={}", resource_limit,
|
|
resource_type, value);
|
|
|
|
const auto type = static_cast<ResourceType>(resource_type);
|
|
if (!IsValidResourceType(type)) {
|
|
LOG_ERROR(Kernel_SVC, "Invalid resource limit type: '{}'", resource_type);
|
|
return ERR_INVALID_ENUM_VALUE;
|
|
}
|
|
|
|
auto& kernel = Core::System::GetInstance().Kernel();
|
|
auto* const current_process = kernel.CurrentProcess();
|
|
ASSERT(current_process != nullptr);
|
|
|
|
auto resource_limit_object =
|
|
current_process->GetHandleTable().Get<ResourceLimit>(resource_limit);
|
|
if (!resource_limit_object) {
|
|
LOG_ERROR(Kernel_SVC, "Handle to non-existent resource limit instance used. Handle={:08X}",
|
|
resource_limit);
|
|
return ERR_INVALID_HANDLE;
|
|
}
|
|
|
|
const auto set_result = resource_limit_object->SetLimitValue(type, static_cast<s64>(value));
|
|
if (set_result.IsError()) {
|
|
LOG_ERROR(
|
|
Kernel_SVC,
|
|
"Attempted to lower resource limit ({}) for category '{}' below its current value ({})",
|
|
resource_limit_object->GetMaxResourceValue(type), resource_type,
|
|
resource_limit_object->GetCurrentResourceValue(type));
|
|
return set_result;
|
|
}
|
|
|
|
return RESULT_SUCCESS;
|
|
}
|
|
|
|
namespace {
|
|
struct FunctionDef {
|
|
using Func = void();
|
|
|
|
u32 id;
|
|
Func* func;
|
|
const char* name;
|
|
};
|
|
} // namespace
|
|
|
|
static const FunctionDef SVC_Table[] = {
|
|
{0x00, nullptr, "Unknown"},
|
|
{0x01, SvcWrap<SetHeapSize>, "SetHeapSize"},
|
|
{0x02, SvcWrap<SetMemoryPermission>, "SetMemoryPermission"},
|
|
{0x03, SvcWrap<SetMemoryAttribute>, "SetMemoryAttribute"},
|
|
{0x04, SvcWrap<MapMemory>, "MapMemory"},
|
|
{0x05, SvcWrap<UnmapMemory>, "UnmapMemory"},
|
|
{0x06, SvcWrap<QueryMemory>, "QueryMemory"},
|
|
{0x07, SvcWrap<ExitProcess>, "ExitProcess"},
|
|
{0x08, SvcWrap<CreateThread>, "CreateThread"},
|
|
{0x09, SvcWrap<StartThread>, "StartThread"},
|
|
{0x0A, SvcWrap<ExitThread>, "ExitThread"},
|
|
{0x0B, SvcWrap<SleepThread>, "SleepThread"},
|
|
{0x0C, SvcWrap<GetThreadPriority>, "GetThreadPriority"},
|
|
{0x0D, SvcWrap<SetThreadPriority>, "SetThreadPriority"},
|
|
{0x0E, SvcWrap<GetThreadCoreMask>, "GetThreadCoreMask"},
|
|
{0x0F, SvcWrap<SetThreadCoreMask>, "SetThreadCoreMask"},
|
|
{0x10, SvcWrap<GetCurrentProcessorNumber>, "GetCurrentProcessorNumber"},
|
|
{0x11, nullptr, "SignalEvent"},
|
|
{0x12, SvcWrap<ClearEvent>, "ClearEvent"},
|
|
{0x13, SvcWrap<MapSharedMemory>, "MapSharedMemory"},
|
|
{0x14, SvcWrap<UnmapSharedMemory>, "UnmapSharedMemory"},
|
|
{0x15, SvcWrap<CreateTransferMemory>, "CreateTransferMemory"},
|
|
{0x16, SvcWrap<CloseHandle>, "CloseHandle"},
|
|
{0x17, SvcWrap<ResetSignal>, "ResetSignal"},
|
|
{0x18, SvcWrap<WaitSynchronization>, "WaitSynchronization"},
|
|
{0x19, SvcWrap<CancelSynchronization>, "CancelSynchronization"},
|
|
{0x1A, SvcWrap<ArbitrateLock>, "ArbitrateLock"},
|
|
{0x1B, SvcWrap<ArbitrateUnlock>, "ArbitrateUnlock"},
|
|
{0x1C, SvcWrap<WaitProcessWideKeyAtomic>, "WaitProcessWideKeyAtomic"},
|
|
{0x1D, SvcWrap<SignalProcessWideKey>, "SignalProcessWideKey"},
|
|
{0x1E, SvcWrap<GetSystemTick>, "GetSystemTick"},
|
|
{0x1F, SvcWrap<ConnectToNamedPort>, "ConnectToNamedPort"},
|
|
{0x20, nullptr, "SendSyncRequestLight"},
|
|
{0x21, SvcWrap<SendSyncRequest>, "SendSyncRequest"},
|
|
{0x22, nullptr, "SendSyncRequestWithUserBuffer"},
|
|
{0x23, nullptr, "SendAsyncRequestWithUserBuffer"},
|
|
{0x24, SvcWrap<GetProcessId>, "GetProcessId"},
|
|
{0x25, SvcWrap<GetThreadId>, "GetThreadId"},
|
|
{0x26, SvcWrap<Break>, "Break"},
|
|
{0x27, SvcWrap<OutputDebugString>, "OutputDebugString"},
|
|
{0x28, nullptr, "ReturnFromException"},
|
|
{0x29, SvcWrap<GetInfo>, "GetInfo"},
|
|
{0x2A, nullptr, "FlushEntireDataCache"},
|
|
{0x2B, nullptr, "FlushDataCache"},
|
|
{0x2C, nullptr, "MapPhysicalMemory"},
|
|
{0x2D, nullptr, "UnmapPhysicalMemory"},
|
|
{0x2E, nullptr, "GetFutureThreadInfo"},
|
|
{0x2F, nullptr, "GetLastThreadInfo"},
|
|
{0x30, SvcWrap<GetResourceLimitLimitValue>, "GetResourceLimitLimitValue"},
|
|
{0x31, SvcWrap<GetResourceLimitCurrentValue>, "GetResourceLimitCurrentValue"},
|
|
{0x32, SvcWrap<SetThreadActivity>, "SetThreadActivity"},
|
|
{0x33, SvcWrap<GetThreadContext>, "GetThreadContext"},
|
|
{0x34, SvcWrap<WaitForAddress>, "WaitForAddress"},
|
|
{0x35, SvcWrap<SignalToAddress>, "SignalToAddress"},
|
|
{0x36, nullptr, "Unknown"},
|
|
{0x37, nullptr, "Unknown"},
|
|
{0x38, nullptr, "Unknown"},
|
|
{0x39, nullptr, "Unknown"},
|
|
{0x3A, nullptr, "Unknown"},
|
|
{0x3B, nullptr, "Unknown"},
|
|
{0x3C, nullptr, "DumpInfo"},
|
|
{0x3D, nullptr, "DumpInfoNew"},
|
|
{0x3E, nullptr, "Unknown"},
|
|
{0x3F, nullptr, "Unknown"},
|
|
{0x40, nullptr, "CreateSession"},
|
|
{0x41, nullptr, "AcceptSession"},
|
|
{0x42, nullptr, "ReplyAndReceiveLight"},
|
|
{0x43, nullptr, "ReplyAndReceive"},
|
|
{0x44, nullptr, "ReplyAndReceiveWithUserBuffer"},
|
|
{0x45, SvcWrap<CreateEvent>, "CreateEvent"},
|
|
{0x46, nullptr, "Unknown"},
|
|
{0x47, nullptr, "Unknown"},
|
|
{0x48, nullptr, "MapPhysicalMemoryUnsafe"},
|
|
{0x49, nullptr, "UnmapPhysicalMemoryUnsafe"},
|
|
{0x4A, nullptr, "SetUnsafeLimit"},
|
|
{0x4B, nullptr, "CreateCodeMemory"},
|
|
{0x4C, nullptr, "ControlCodeMemory"},
|
|
{0x4D, nullptr, "SleepSystem"},
|
|
{0x4E, nullptr, "ReadWriteRegister"},
|
|
{0x4F, nullptr, "SetProcessActivity"},
|
|
{0x50, SvcWrap<CreateSharedMemory>, "CreateSharedMemory"},
|
|
{0x51, nullptr, "MapTransferMemory"},
|
|
{0x52, nullptr, "UnmapTransferMemory"},
|
|
{0x53, nullptr, "CreateInterruptEvent"},
|
|
{0x54, nullptr, "QueryPhysicalAddress"},
|
|
{0x55, nullptr, "QueryIoMapping"},
|
|
{0x56, nullptr, "CreateDeviceAddressSpace"},
|
|
{0x57, nullptr, "AttachDeviceAddressSpace"},
|
|
{0x58, nullptr, "DetachDeviceAddressSpace"},
|
|
{0x59, nullptr, "MapDeviceAddressSpaceByForce"},
|
|
{0x5A, nullptr, "MapDeviceAddressSpaceAligned"},
|
|
{0x5B, nullptr, "MapDeviceAddressSpace"},
|
|
{0x5C, nullptr, "UnmapDeviceAddressSpace"},
|
|
{0x5D, nullptr, "InvalidateProcessDataCache"},
|
|
{0x5E, nullptr, "StoreProcessDataCache"},
|
|
{0x5F, nullptr, "FlushProcessDataCache"},
|
|
{0x60, nullptr, "DebugActiveProcess"},
|
|
{0x61, nullptr, "BreakDebugProcess"},
|
|
{0x62, nullptr, "TerminateDebugProcess"},
|
|
{0x63, nullptr, "GetDebugEvent"},
|
|
{0x64, nullptr, "ContinueDebugEvent"},
|
|
{0x65, nullptr, "GetProcessList"},
|
|
{0x66, nullptr, "GetThreadList"},
|
|
{0x67, nullptr, "GetDebugThreadContext"},
|
|
{0x68, nullptr, "SetDebugThreadContext"},
|
|
{0x69, nullptr, "QueryDebugProcessMemory"},
|
|
{0x6A, nullptr, "ReadDebugProcessMemory"},
|
|
{0x6B, nullptr, "WriteDebugProcessMemory"},
|
|
{0x6C, nullptr, "SetHardwareBreakPoint"},
|
|
{0x6D, nullptr, "GetDebugThreadParam"},
|
|
{0x6E, nullptr, "Unknown"},
|
|
{0x6F, nullptr, "GetSystemInfo"},
|
|
{0x70, nullptr, "CreatePort"},
|
|
{0x71, nullptr, "ManageNamedPort"},
|
|
{0x72, nullptr, "ConnectToPort"},
|
|
{0x73, nullptr, "SetProcessMemoryPermission"},
|
|
{0x74, nullptr, "MapProcessMemory"},
|
|
{0x75, nullptr, "UnmapProcessMemory"},
|
|
{0x76, nullptr, "QueryProcessMemory"},
|
|
{0x77, nullptr, "MapProcessCodeMemory"},
|
|
{0x78, nullptr, "UnmapProcessCodeMemory"},
|
|
{0x79, nullptr, "CreateProcess"},
|
|
{0x7A, nullptr, "StartProcess"},
|
|
{0x7B, nullptr, "TerminateProcess"},
|
|
{0x7C, SvcWrap<GetProcessInfo>, "GetProcessInfo"},
|
|
{0x7D, SvcWrap<CreateResourceLimit>, "CreateResourceLimit"},
|
|
{0x7E, SvcWrap<SetResourceLimitLimitValue>, "SetResourceLimitLimitValue"},
|
|
{0x7F, nullptr, "CallSecureMonitor"},
|
|
};
|
|
|
|
static const FunctionDef* GetSVCInfo(u32 func_num) {
|
|
if (func_num >= std::size(SVC_Table)) {
|
|
LOG_ERROR(Kernel_SVC, "Unknown svc=0x{:02X}", func_num);
|
|
return nullptr;
|
|
}
|
|
return &SVC_Table[func_num];
|
|
}
|
|
|
|
MICROPROFILE_DEFINE(Kernel_SVC, "Kernel", "SVC", MP_RGB(70, 200, 70));
|
|
|
|
void CallSVC(u32 immediate) {
|
|
MICROPROFILE_SCOPE(Kernel_SVC);
|
|
|
|
// Lock the global kernel mutex when we enter the kernel HLE.
|
|
std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
|
|
|
|
const FunctionDef* info = GetSVCInfo(immediate);
|
|
if (info) {
|
|
if (info->func) {
|
|
info->func();
|
|
} else {
|
|
LOG_CRITICAL(Kernel_SVC, "Unimplemented SVC function {}(..)", info->name);
|
|
}
|
|
} else {
|
|
LOG_CRITICAL(Kernel_SVC, "Unknown SVC function 0x{:X}", immediate);
|
|
}
|
|
}
|
|
|
|
} // namespace Kernel
|