ladybird/Kernel/Arch/x86/init.cpp

416 lines
16 KiB
C++
Raw Normal View History

/*
* Copyright (c) 2018-2020, Andreas Kling <kling@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#include <AK/Types.h>
#include <Kernel/Arch/InterruptManagement.h>
#include <Kernel/Arch/Processor.h>
#include <Kernel/Arch/x86/Hypervisor/VMWareBackdoor.h>
#include <Kernel/Arch/x86/common/Interrupts/APIC.h>
#include <Kernel/Arch/x86/common/Interrupts/PIC.h>
#include <Kernel/BootInfo.h>
#include <Kernel/Bus/PCI/Access.h>
#include <Kernel/Bus/PCI/Initializer.h>
#include <Kernel/Bus/USB/USBManagement.h>
#include <Kernel/Bus/VirtIO/Device.h>
#include <Kernel/CommandLine.h>
#include <Kernel/Devices/Audio/Management.h>
#include <Kernel/Devices/DeviceControlDevice.h>
#include <Kernel/Devices/DeviceManagement.h>
#include <Kernel/Devices/FullDevice.h>
#include <Kernel/Devices/HID/HIDManagement.h>
#include <Kernel/Devices/KCOVDevice.h>
#include <Kernel/Devices/MemoryDevice.h>
#include <Kernel/Devices/NullDevice.h>
#include <Kernel/Devices/PCISerialDevice.h>
#include <Kernel/Devices/RandomDevice.h>
#include <Kernel/Devices/SelfTTYDevice.h>
#include <Kernel/Devices/SerialDevice.h>
#include <Kernel/Devices/ZeroDevice.h>
#include <Kernel/FileSystem/SysFS/Registry.h>
#include <Kernel/FileSystem/SysFS/Subsystems/Firmware/Directory.h>
#include <Kernel/FileSystem/VirtualFileSystem.h>
#include <Kernel/Firmware/ACPI/Initialize.h>
#include <Kernel/Firmware/ACPI/Parser.h>
#include <Kernel/Graphics/Console/BootFramebufferConsole.h>
#include <Kernel/Graphics/Console/VGATextModeConsole.h>
#include <Kernel/Graphics/GraphicsManagement.h>
#include <Kernel/Heap/kmalloc.h>
#include <Kernel/JailManagement.h>
#include <Kernel/KSyms.h>
#include <Kernel/Memory/MemoryManager.h>
#include <Kernel/Multiboot.h>
#include <Kernel/Net/NetworkTask.h>
#include <Kernel/Net/NetworkingManagement.h>
#include <Kernel/Panic.h>
#include <Kernel/Prekernel/Prekernel.h>
#include <Kernel/Process.h>
Kernel: Introduce the new ProcFS design The new ProcFS design consists of two main parts: 1. The representative ProcFS class, which is derived from the FS class. The ProcFS and its inodes are much more lean - merely 3 classes to represent the common type of inodes - regular files, symbolic links and directories. They're backed by a ProcFSExposedComponent object, which is responsible for the functional operation behind the scenes. 2. The backend of the ProcFS - the ProcFSComponentsRegistrar class and all derived classes from the ProcFSExposedComponent class. These together form the entire backend and handle all the functions you can expect from the ProcFS. The ProcFSExposedComponent derived classes split to 3 types in the manner of lifetime in the kernel: 1. Persistent objects - this category includes all basic objects, like the root folder, /proc/bus folder, main blob files in the root folders, etc. These objects are persistent and cannot die ever. 2. Semi-persistent objects - this category includes all PID folders, and subdirectories to the PID folders. It also includes exposed objects like the unveil JSON'ed blob. These object are persistent as long as the the responsible process they represent is still alive. 3. Dynamic objects - this category includes files in the subdirectories of a PID folder, like /proc/PID/fd/* or /proc/PID/stacks/*. Essentially, these objects are always created dynamically and when no longer in need after being used, they're deallocated. Nevertheless, the new allocated backend objects and inodes try to use the same InodeIndex if possible - this might change only when a thread dies and a new thread is born with a new thread stack, or when a file descriptor is closed and a new one within the same file descriptor number is opened. This is needed to actually be able to do something useful with these objects. The new design assures that many ProcFS instances can be used at once, with one backend for usage for all instances.
2021-06-12 04:23:58 +03:00
#include <Kernel/ProcessExposed.h>
#include <Kernel/Random.h>
#include <Kernel/Scheduler.h>
#include <Kernel/Sections.h>
#include <Kernel/Storage/StorageManagement.h>
#include <Kernel/TTY/ConsoleManagement.h>
#include <Kernel/TTY/PTYMultiplexer.h>
#include <Kernel/TTY/VirtualConsole.h>
#include <Kernel/Tasks/FinalizerTask.h>
#include <Kernel/Tasks/SyncTask.h>
Kernel: Introduce the new Time management subsystem This new subsystem includes better abstractions of how time will be handled in the OS. We take advantage of the existing RTC timer to aid in keeping time synchronized. This is standing in contrast to how we handled time-keeping in the kernel, where the PIT was responsible for that function in addition to update the scheduler about ticks. With that new advantage, we can easily change the ticking dynamically and still keep the time synchronized. In the process context, we no longer use a fixed declaration of TICKS_PER_SECOND, but we call the TimeManagement singleton class to provide us the right value. This allows us to use dynamic ticking in the future, a feature known as tickless kernel. The scheduler no longer does by himself the calculation of real time (Unix time), and just calls the TimeManagment singleton class to provide the value. Also, we can use 2 new boot arguments: - the "time" boot argument accpets either the value "modern", or "legacy". If "modern" is specified, the time management subsystem will try to setup HPET. Otherwise, for "legacy" value, the time subsystem will revert to use the PIT & RTC, leaving HPET disabled. If this boot argument is not specified, the default pattern is to try to setup HPET. - the "hpet" boot argumet accepts either the value "periodic" or "nonperiodic". If "periodic" is specified, the HPET will scan for periodic timers, and will assert if none are found. If only one is found, that timer will be assigned for the time-keeping task. If more than one is found, both time-keeping task & scheduler-ticking task will be assigned to periodic timers. If this boot argument is not specified, the default pattern is to try to scan for HPET periodic timers. This boot argument has no effect if HPET is disabled. In hardware context, PIT & RealTimeClock classes are merely inheriting from the HardwareTimer class, and they allow to use the old i8254 (PIT) and RTC devices, managing them via IO ports. By default, the RTC will be programmed to a frequency of 1024Hz. The PIT will be programmed to a frequency close to 1000Hz. About HPET, depending if we need to scan for periodic timers or not, we try to set a frequency close to 1000Hz for the time-keeping timer and scheduler-ticking timer. Also, if possible, we try to enable the Legacy replacement feature of the HPET. This feature if exists, instructs the chipset to disconnect both i8254 (PIT) and RTC. This behavior is observable on QEMU, and was verified against the source code: https://github.com/qemu/qemu/commit/ce967e2f33861b0e17753f97fa4527b5943c94b6 The HPETComparator class is inheriting from HardwareTimer class, and is responsible for an individual HPET comparator, which is essentially a timer. Therefore, it needs to call the singleton HPET class to perform HPET-related operations. The new abstraction of Hardware timers brings an opportunity of more new features in the foreseeable future. For example, we can change the callback function of each hardware timer, thus it makes it possible to swap missions between hardware timers, or to allow to use a hardware timer for other temporary missions (e.g. calibrating the LAPIC timer, measuring the CPU frequency, etc).
2020-03-09 18:03:27 +03:00
#include <Kernel/Time/TimeManagement.h>
#include <Kernel/WorkQueue.h>
#include <Kernel/kstdio.h>
// Defined in the linker script
typedef void (*ctor_func_t)();
extern ctor_func_t start_heap_ctors[];
extern ctor_func_t end_heap_ctors[];
extern ctor_func_t start_ctors[];
extern ctor_func_t end_ctors[];
extern uintptr_t __stack_chk_guard;
READONLY_AFTER_INIT uintptr_t __stack_chk_guard __attribute__((used));
extern "C" u8 start_of_safemem_text[];
extern "C" u8 end_of_safemem_text[];
extern "C" u8 start_of_safemem_atomic_text[];
extern "C" u8 end_of_safemem_atomic_text[];
extern "C" u8 end_of_kernel_image[];
multiboot_module_entry_t multiboot_copy_boot_modules_array[16];
size_t multiboot_copy_boot_modules_count;
READONLY_AFTER_INIT bool g_in_early_boot;
namespace Kernel {
[[noreturn]] static void init_stage2(void*);
static void setup_serial_debug();
2021-03-03 16:16:14 +03:00
// boot.S expects these functions to exactly have the following signatures.
// We declare them here to ensure their signatures don't accidentally change.
extern "C" void init_finished(u32 cpu) __attribute__((used));
extern "C" [[noreturn]] void init_ap(FlatPtr cpu, Processor* processor_info);
extern "C" [[noreturn]] void init(BootInfo const&);
READONLY_AFTER_INIT VirtualConsole* tty0;
ProcessID g_init_pid { 0 };
ALWAYS_INLINE static Processor& bsp_processor()
{
// This solves a problem where the bsp Processor instance
// gets "re"-initialized in init() when we run all global constructors.
alignas(Processor) static u8 bsp_processor_storage[sizeof(Processor)];
return (Processor&)bsp_processor_storage;
}
// SerenityOS Kernel C++ entry point :^)
//
// This is where C++ execution begins, after boot.S transfers control here.
//
// The purpose of init() is to start multi-tasking. It does the bare minimum
// amount of work needed to start the scheduler.
//
// Once multi-tasking is ready, we spawn a new thread that starts in the
// init_stage2() function. Initialization continues there.
extern "C" {
READONLY_AFTER_INIT PhysicalAddress start_of_prekernel_image;
READONLY_AFTER_INIT PhysicalAddress end_of_prekernel_image;
READONLY_AFTER_INIT size_t physical_to_virtual_offset;
READONLY_AFTER_INIT FlatPtr kernel_mapping_base;
READONLY_AFTER_INIT FlatPtr kernel_load_base;
READONLY_AFTER_INIT PhysicalAddress boot_pml4t;
READONLY_AFTER_INIT PhysicalAddress boot_pdpt;
READONLY_AFTER_INIT PhysicalAddress boot_pd0;
READONLY_AFTER_INIT PhysicalAddress boot_pd_kernel;
READONLY_AFTER_INIT PageTableEntry* boot_pd_kernel_pt1023;
2022-04-01 20:58:27 +03:00
READONLY_AFTER_INIT char const* kernel_cmdline;
READONLY_AFTER_INIT u32 multiboot_flags;
READONLY_AFTER_INIT multiboot_memory_map_t* multiboot_memory_map;
READONLY_AFTER_INIT size_t multiboot_memory_map_count;
READONLY_AFTER_INIT multiboot_module_entry_t* multiboot_modules;
READONLY_AFTER_INIT size_t multiboot_modules_count;
READONLY_AFTER_INIT PhysicalAddress multiboot_framebuffer_addr;
READONLY_AFTER_INIT u32 multiboot_framebuffer_pitch;
READONLY_AFTER_INIT u32 multiboot_framebuffer_width;
READONLY_AFTER_INIT u32 multiboot_framebuffer_height;
READONLY_AFTER_INIT u8 multiboot_framebuffer_bpp;
READONLY_AFTER_INIT u8 multiboot_framebuffer_type;
}
Atomic<Graphics::Console*> g_boot_console;
extern "C" [[noreturn]] UNMAP_AFTER_INIT void init(BootInfo const& boot_info)
{
g_in_early_boot = true;
start_of_prekernel_image = PhysicalAddress { boot_info.start_of_prekernel_image };
end_of_prekernel_image = PhysicalAddress { boot_info.end_of_prekernel_image };
physical_to_virtual_offset = boot_info.physical_to_virtual_offset;
kernel_mapping_base = boot_info.kernel_mapping_base;
kernel_load_base = boot_info.kernel_load_base;
gdt64ptr = boot_info.gdt64ptr;
code64_sel = boot_info.code64_sel;
boot_pml4t = PhysicalAddress { boot_info.boot_pml4t };
boot_pdpt = PhysicalAddress { boot_info.boot_pdpt };
boot_pd0 = PhysicalAddress { boot_info.boot_pd0 };
boot_pd_kernel = PhysicalAddress { boot_info.boot_pd_kernel };
boot_pd_kernel_pt1023 = (PageTableEntry*)boot_info.boot_pd_kernel_pt1023;
kernel_cmdline = (char const*)boot_info.kernel_cmdline;
multiboot_flags = boot_info.multiboot_flags;
multiboot_memory_map = (multiboot_memory_map_t*)boot_info.multiboot_memory_map;
multiboot_memory_map_count = boot_info.multiboot_memory_map_count;
multiboot_modules = (multiboot_module_entry_t*)boot_info.multiboot_modules;
multiboot_modules_count = boot_info.multiboot_modules_count;
multiboot_framebuffer_addr = PhysicalAddress { boot_info.multiboot_framebuffer_addr };
multiboot_framebuffer_pitch = boot_info.multiboot_framebuffer_pitch;
multiboot_framebuffer_width = boot_info.multiboot_framebuffer_width;
multiboot_framebuffer_height = boot_info.multiboot_framebuffer_height;
multiboot_framebuffer_bpp = boot_info.multiboot_framebuffer_bpp;
multiboot_framebuffer_type = boot_info.multiboot_framebuffer_type;
setup_serial_debug();
// We need to copy the command line before kmalloc is initialized,
// as it may overwrite parts of multiboot!
CommandLine::early_initialize(kernel_cmdline);
memcpy(multiboot_copy_boot_modules_array, multiboot_modules, multiboot_modules_count * sizeof(multiboot_module_entry_t));
multiboot_copy_boot_modules_count = multiboot_modules_count;
new (&bsp_processor()) Processor();
bsp_processor().early_initialize(0);
// Invoke the constructors needed for the kernel heap
for (ctor_func_t* ctor = start_heap_ctors; ctor < end_heap_ctors; ctor++)
(*ctor)();
kmalloc_init();
load_kernel_symbol_table();
bsp_processor().initialize(0);
CommandLine::initialize();
Memory::MemoryManager::initialize(0);
// NOTE: If the bootloader provided a framebuffer, then set up an initial console.
// If the bootloader didn't provide a framebuffer, then set up an initial text console.
// We do so we can see the output on the screen as soon as possible.
if (!kernel_command_line().is_early_boot_console_disabled()) {
if (!multiboot_framebuffer_addr.is_null() && multiboot_framebuffer_type == MULTIBOOT_FRAMEBUFFER_TYPE_RGB) {
g_boot_console = &try_make_lock_ref_counted<Graphics::BootFramebufferConsole>(multiboot_framebuffer_addr, multiboot_framebuffer_width, multiboot_framebuffer_height, multiboot_framebuffer_pitch).value().leak_ref();
} else {
g_boot_console = &Graphics::VGATextModeConsole::initialize().leak_ref();
}
}
dmesgln("Starting SerenityOS...");
DeviceManagement::initialize();
SysFSComponentRegistry::initialize();
DeviceManagement::the().attach_null_device(*NullDevice::must_initialize());
DeviceManagement::the().attach_console_device(*ConsoleDevice::must_create());
DeviceManagement::the().attach_device_control_device(*DeviceControlDevice::must_create());
MM.unmap_prekernel();
// Ensure that the safemem sections are not empty. This could happen if the linker accidentally discards the sections.
VERIFY(+start_of_safemem_text != +end_of_safemem_text);
VERIFY(+start_of_safemem_atomic_text != +end_of_safemem_atomic_text);
// Invoke all static global constructors in the kernel.
// Note that we want to do this as early as possible.
for (ctor_func_t* ctor = start_ctors; ctor < end_ctors; ctor++)
(*ctor)();
InterruptManagement::initialize();
ACPI::initialize();
// Initialize TimeManagement before using randomness!
TimeManagement::initialize(0);
__stack_chk_guard = get_fast_random<uintptr_t>();
ProcFSComponentRegistry::initialize();
JailManagement::the();
Process::initialize();
Scheduler::initialize();
if (APIC::initialized() && APIC::the().enabled_processor_count() > 1) {
// We must set up the AP boot environment before switching to a kernel process,
// as pages below address USER_RANGE_BASE are only accessible through the kernel
// page directory.
APIC::the().setup_ap_boot_environment();
}
{
LockRefPtr<Thread> init_stage2_thread;
(void)Process::create_kernel_process(init_stage2_thread, KString::must_create("init_stage2"sv), init_stage2, nullptr, THREAD_AFFINITY_DEFAULT, Process::RegisterProcess::No);
// We need to make sure we drop the reference for init_stage2_thread
// before calling into Scheduler::start, otherwise we will have a
// dangling Thread that never gets cleaned up
}
Scheduler::start();
VERIFY_NOT_REACHED();
}
//
// This is where C++ execution begins for APs, after boot.S transfers control here.
//
// The purpose of init_ap() is to initialize APs for multi-tasking.
//
extern "C" [[noreturn]] UNMAP_AFTER_INIT void init_ap(FlatPtr cpu, Processor* processor_info)
{
processor_info->early_initialize(cpu);
processor_info->initialize(cpu);
Memory::MemoryManager::initialize(cpu);
Scheduler::set_idle_thread(APIC::the().get_idle_thread(cpu));
Scheduler::start();
VERIFY_NOT_REACHED();
}
//
// This method is called once a CPU enters the scheduler and its idle thread
// At this point the initial boot stack can be freed
//
extern "C" UNMAP_AFTER_INIT void init_finished(u32 cpu)
{
if (cpu == 0) {
// TODO: we can reuse the boot stack, maybe for kmalloc()?
} else {
APIC::the().init_finished(cpu);
TimeManagement::initialize(cpu);
}
}
void init_stage2(void*)
{
// This is a little bit of a hack. We can't register our process at the time we're
// creating it, but we need to be registered otherwise finalization won't be happy.
// The colonel process gets away without having to do this because it never exits.
Process::register_new(Process::current());
WorkQueue::initialize();
if (kernel_command_line().is_smp_enabled() && APIC::initialized() && APIC::the().enabled_processor_count() > 1) {
// We can't start the APs until we have a scheduler up and running.
// We need to be able to process ICI messages, otherwise another
// core may send too many and end up deadlocking once the pool is
// exhausted
APIC::the().boot_aps();
}
// Initialize the PCI Bus as early as possible, for early boot (PCI based) serial logging
PCI::initialize();
if (!PCI::Access::is_disabled()) {
PCISerialDevice::detect();
}
VirtualFileSystem::initialize();
if (!is_serial_debug_enabled())
(void)SerialDevice::must_create(0).leak_ref();
(void)SerialDevice::must_create(1).leak_ref();
(void)SerialDevice::must_create(2).leak_ref();
(void)SerialDevice::must_create(3).leak_ref();
VMWareBackdoor::the(); // don't wait until first mouse packet
MUST(HIDManagement::initialize());
2021-06-12 16:07:44 +03:00
GraphicsManagement::the().initialize();
ConsoleManagement::the().initialize();
SyncTask::spawn();
FinalizerTask::spawn();
auto boot_profiling = kernel_command_line().is_boot_profiling_enabled();
if (!PCI::Access::is_disabled()) {
USB::USBManagement::initialize();
}
FirmwareSysFSDirectory::initialize();
if (!PCI::Access::is_disabled()) {
VirtIO::detect();
}
NetworkingManagement::the().initialize();
Syscall::initialize();
#ifdef ENABLE_KERNEL_COVERAGE_COLLECTION
(void)KCOVDevice::must_create().leak_ref();
#endif
(void)MemoryDevice::must_create().leak_ref();
(void)ZeroDevice::must_create().leak_ref();
(void)FullDevice::must_create().leak_ref();
(void)RandomDevice::must_create().leak_ref();
(void)SelfTTYDevice::must_create().leak_ref();
PTYMultiplexer::initialize();
AudioManagement::the().initialize();
StorageManagement::the().initialize(kernel_command_line().root_device(), kernel_command_line().is_force_pio(), kernel_command_line().is_nvme_polling_enabled());
if (VirtualFileSystem::the().mount_root(StorageManagement::the().root_filesystem()).is_error()) {
PANIC("VirtualFileSystem::mount_root failed");
}
2018-10-17 12:44:06 +03:00
// Switch out of early boot mode.
g_in_early_boot = false;
// NOTE: Everything marked READONLY_AFTER_INIT becomes non-writable after this point.
MM.protect_readonly_after_init_memory();
// NOTE: Everything in the .ksyms section becomes read-only after this point.
MM.protect_ksyms_after_init();
// NOTE: Everything marked UNMAP_AFTER_INIT becomes inaccessible after this point.
MM.unmap_text_after_init();
LockRefPtr<Thread> thread;
auto userspace_init = kernel_command_line().userspace_init();
auto init_args = kernel_command_line().userspace_init_args();
auto init_or_error = Process::try_create_user_process(thread, userspace_init, UserID(0), GroupID(0), move(init_args), {}, tty0);
if (init_or_error.is_error())
PANIC("init_stage2: Error spawning init process: {}", init_or_error.error());
g_init_pid = init_or_error.value()->pid();
thread->set_priority(THREAD_PRIORITY_HIGH);
if (boot_profiling) {
dbgln("Starting full system boot profiling");
MutexLocker mutex_locker(Process::current().big_lock());
2022-04-01 20:58:27 +03:00
auto const enable_all = ~(u64)0;
auto result = Process::current().profiling_enable(-1, enable_all);
VERIFY(!result.is_error());
}
NetworkTask::spawn();
Process::current().sys$exit(0);
VERIFY_NOT_REACHED();
}
UNMAP_AFTER_INIT void setup_serial_debug()
{
// serial_debug will output all the dbgln() data to COM1 at
// 8-N-1 57600 baud. this is particularly useful for debugging the boot
// process on live hardware.
if (StringView { kernel_cmdline, strlen(kernel_cmdline) }.contains("serial_debug"sv)) {
set_serial_debug_enabled(true);
}
}
// Define some Itanium C++ ABI methods to stop the linker from complaining.
// If we actually call these something has gone horribly wrong
void* __dso_handle __attribute__((visibility("hidden")));
}