Merge pull request #170 from assembler-0/Development

Development

Author: assembler-0

CMakeLists.txt

@@ -217,7 +217,7 @@ add_custom_target(extra-img
         COMMAND ${QEMU_IMG} create -f raw VirtioDisk.img 128M
         COMMAND ${MKFS_EXT2} VirtioDisk.img
         COMMAND ${QEMU_IMG} create -f raw SataDisk.img 128M
-        COMMAND ${MKFS_EXT2} SataDisk.img
+        COMMAND ${MKFS_FAT} -F12 SataDisk.img
         COMMAND ${QEMU_IMG} create -f raw NVMeDisk.img 256M
         COMMAND ${MKFS_EXT2} NVMeDisk.img
         COMMENT "Creating extra disk images"

README.md

@@ -1,10 +1,10 @@
 # [VoidFrame] - a ring 0 kernel 💫
 
-> A fast, simple, secure 64-bit ring-0 kernel written in C and assembly. With modern capabilities.
+> A fast, simple, secure 64-bit ring-0 kernel written in C (Rust) and assembly. With modern capabilities.
 
 ---
 
-- How it works: [here!](docs/ARCHITECTURE.md)
+- How it works (outdated): [here!](docs/ARCHITECTURE.md)
 - Development Guide: [here!](docs/DEVELOPMENT.md)
 
 ---
@@ -28,7 +28,6 @@ It would be amazing if you could contribute to this project!
 - x64-compatible cpu (used: Intel i3-12100F)
 - POSIX-compliant OS (SysV ABI) (used: Arch Linux 6.16.9-arch1-1)
 - cmake >= 3.20 (used: cmake 4.1.2)
-- meson >= 1.4 (used: meson 1.9.1)
 - ninja >= 1.11 (used: ninja 1.21.1)
 - clang/++ >= 18.0.0 (used: 20.1.8)
 - rustup (nightly, bare metal toolchain) >= 1.89.0 (used: 1.92.0-nightly)
@@ -40,18 +39,6 @@ It would be amazing if you could contribute to this project!
     - Note: depending on your distro, grub-mkrescue may require xorriso and mtools packages.
 
 ### Quickstart
-#### Full development setup
-```bash
-# Meson
-git clone https://github.com/assembler-0/VoidFrame.git
-cd VoidFrame
-meson setup build
-cd build
-ninja -j$(nproc)
-ninja img
-ninja extra-img
-ninja run
-```
 ```bash
 # CMake
 git clone https://github.com/assembler-0/VoidFrame.git
@@ -61,7 +48,7 @@ cd build
 cmake .. -DCMAKE_BUILD_TYPE=Release \
     -DCMAKE_TOOLCHAIN_FILE=../cmake/toolchain/linux-x64.cmake \
     -G Ninja \
-    -DVF_SCHEDULER=<MLFQ/EEVDF>
+    -DVF_SCHEDULER=EEVDF
 ccmake . # Optinal, tune as needed
 ninja -j$(nproc)
 ninja run

drivers/ACPI.c

@@ -2,9 +2,11 @@
 #include "Console.h"
 #include "Io.h"
 #include "MemOps.h"
+#include "Scheduler.h"
 #include "StringOps.h"
 #include "VMem.h"
 #include "TSC.h"
+#include "VFS.h"
 
 static ACPIFADT* g_fadt = NULL;
 static bool g_acpi_initialized = false;
@@ -47,128 +49,126 @@ static bool ValidateChecksum(void* table, uint32_t length) {
 
 // Map ACPI table to virtual memory
 static void* MapACPITable(uint32_t phys_addr, uint32_t size) {
-    // Align to page boundaries
-    uint32_t aligned_addr = phys_addr & ~0xFFF;
-    uint32_t offset = phys_addr - aligned_addr;
-    uint32_t aligned_size = ((size + offset + 0xFFF) & ~0xFFF);
-    
+    uint64_t aligned_addr = phys_addr & ~0xFFF;
+    uint64_t offset = phys_addr - aligned_addr;
+    uint64_t aligned_size = ((size + offset + 0xFFF) & ~0xFFF);
+
     void* virt_addr = VMemAlloc(aligned_size);
-    if (!virt_addr) return NULL;
-    
+    if (!virt_addr) {
+        PrintKernelError("ACPI: Failed to allocate virtual memory for ACPI table\n");
+        return NULL;
+    }
+
     if (VMemUnmap((uint64_t)virt_addr, aligned_size) != VMEM_SUCCESS) {
         VMemFree(virt_addr, aligned_size);
+        PrintKernelError("ACPI: Failed to unmap virtual memory for ACPI table\n");
         return NULL;
     }
-    
+
     if (VMemMapMMIO((uint64_t)virt_addr, aligned_addr, aligned_size, PAGE_WRITABLE | PAGE_NOCACHE) != VMEM_SUCCESS) {
-        VMemFree(virt_addr, aligned_size);
+        // No need to free here as VMemMapMMIO should handle cleanup on failure
+        PrintKernelError("ACPI: Failed to map MMIO for ACPI table\n");
         return NULL;
     }
-    
+
     return (uint8_t*)virt_addr + offset;
 }
 
+static ACPIRSDT* g_rsdt = NULL;
+
+void* AcpiFindTable(const char* signature) {
+    if (!g_rsdt) {
+        return NULL;
+    }
+
+    uint32_t entries = (g_rsdt->header.length - sizeof(ACPISDTHeader)) / 4;
+    for (uint32_t i = 0; i < entries; i++) {
+        ACPISDTHeader* header = (ACPISDTHeader*)MapACPITable(g_rsdt->table_pointers[i], sizeof(ACPISDTHeader));
+        if (!header) {
+            continue;
+        }
+
+        if (FastMemcmp(header->signature, signature, 4) == 0) {
+            void* table = MapACPITable(g_rsdt->table_pointers[i], header->length);
+            VMemUnmap((uint64_t)header - (g_rsdt->table_pointers[i] & 0xFFF), ((sizeof(ACPISDTHeader) + (g_rsdt->table_pointers[i] & 0xFFF) + 0xFFF) & ~0xFFF));
+            return table;
+        }
+
+        VMemUnmap((uint64_t)header - (g_rsdt->table_pointers[i] & 0xFFF), ((sizeof(ACPISDTHeader) + (g_rsdt->table_pointers[i] & 0xFFF) + 0xFFF) & ~0xFFF));
+    }
+
+    return NULL;
+}
+
 bool ACPIInit(void) {
     PrintKernel("ACPI: Initializing ACPI subsystem...\n");
-    
-    // Find RSDP
+
     ACPIRSDPv1* rsdp = FindRSDP();
     if (!rsdp) {
         PrintKernelError("ACPI: RSDP not found\n");
         return false;
     }
-    
-    PrintKernel("ACPI: Found RSDP at 0x");
-    PrintKernelHex((uint64_t)rsdp);
-    PrintKernel("\n");
-    
-    // Validate RSDP checksum
+
     if (!ValidateChecksum(rsdp, sizeof(ACPIRSDPv1))) {
         PrintKernelError("ACPI: Invalid RSDP checksum\n");
         return false;
     }
 
-    ACPIRSDT* rsdt = (ACPIRSDT*)MapACPITable(rsdp->rsdt_address,
-                                             sizeof(ACPISDTHeader));
-    if (!rsdt) {
-        PrintKernelError("ACPI: Failed to map RSDT\n");
+    g_rsdt = (ACPIRSDT*)MapACPITable(rsdp->rsdt_address, sizeof(ACPISDTHeader));
+    if (!g_rsdt) {
+        PrintKernelError("ACPI: Failed to map RSDT header\n");
         return false;
     }
-    // Validate RSDT signature
-    if (FastMemcmp(rsdt->header.signature, ACPI_RSDT_SIG, 4) != 0) {
+
+    if (FastMemcmp(g_rsdt->header.signature, ACPI_RSDT_SIG, 4) != 0) {
         PrintKernelError("ACPI: Invalid RSDT signature\n");
-        // Clean up the initial header‐only mapping
-        VMemUnmap((uint64_t)rsdt - (rsdp->rsdt_address & 0xFFF),
-                  ((sizeof(ACPISDTHeader) + (rsdp->rsdt_address & 0xFFF) + 0xFFF)
-                   & ~0xFFF));
+        VMemUnmap((uint64_t)g_rsdt - (rsdp->rsdt_address & 0xFFF), ((sizeof(ACPISDTHeader) + (rsdp->rsdt_address & 0xFFF) + 0xFFF) & ~0xFFF));
         return false;
     }
-    // Remap with the full table length
-    uint32_t rsdt_size = rsdt->header.length;
-    // Remember the old header mapping so we can free it afterward
-    void*    old_rsdt   = rsdt;
-    uint32_t old_offset = rsdp->rsdt_address & 0xFFF;
-    rsdt = (ACPIRSDT*)MapACPITable(rsdp->rsdt_address, rsdt_size);
-    // Now unmap the temporary header‐only region
-    VMemUnmap((uint64_t)old_rsdt - old_offset,
-              ((sizeof(ACPISDTHeader) + old_offset + 0xFFF) & ~0xFFF));
-    
-    PrintKernel("ACPI: RSDT mapped, length=");
-    PrintKernelInt(rsdt_size);
-    PrintKernel("\n");
-    
-    // Find FADT
-    uint32_t entries = (rsdt_size - sizeof(ACPISDTHeader)) / 4;
-    for (uint32_t i = 0; i < entries; i++) {
-        ACPISDTHeader* header = (ACPISDTHeader*)MapACPITable(
-            rsdt->table_pointers[i],
-            sizeof(ACPISDTHeader)
-        );
-        if (!header) continue;
-
-        bool is_fadt = FastMemcmp(header->signature, ACPI_FADT_SIG, 4) == 0;
-        uint32_t table_length = header->length;
-        uint32_t header_offset = rsdt->table_pointers[i] & 0xFFF;
-
-        // Unmap the header mapping now that we've inspected it
-        VMemUnmap(
-            (uint64_t)header - header_offset,
-            ( (sizeof(ACPISDTHeader) + header_offset + 0xFFF) & ~0xFFF )
-        );
-
-        if (is_fadt) {
-            PrintKernel("ACPI: Found FADT\n");
-            g_fadt = (ACPIFADT*)MapACPITable(
-                rsdt->table_pointers[i],
-                table_length
-            );
-            break;
-        }
+
+    uint32_t rsdt_size = g_rsdt->header.length;
+    VMemUnmap((uint64_t)g_rsdt - (rsdp->rsdt_address & 0xFFF), ((sizeof(ACPISDTHeader) + (rsdp->rsdt_address & 0xFFF) + 0xFFF) & ~0xFFF));
+    g_rsdt = (ACPIRSDT*)MapACPITable(rsdp->rsdt_address, rsdt_size);
+    if (!g_rsdt) {
+        PrintKernelError("ACPI: Failed to map full RSDT\n");
+        return false;
     }
-    
-    PrintKernelError("ACPI: FADT not found or invalid\n");
-    return false;
+
+    g_fadt = (ACPIFADT*)AcpiFindTable(ACPI_FADT_SIG);
+    if (!g_fadt) {
+        PrintKernelError("ACPI: FADT not found or invalid\n");
+        return false;
+    }
+
+    g_acpi_initialized = true;
+    PrintKernelSuccess("ACPI: Subsystem initialized\n");
+    return true;
+}
+
+void ACPIResetProcedure() {
+    PrintKernel("ACPI: Unmounting Filesystems...\n");
+    VfsUnmountAll();
+    PrintKernelSuccess("ACPI: Filesystems unmounted\n");
+
+    PrintKernel("ACPI: Stopping all processes and services...\n");
+    KillAllProcess("SHUTDOWN");
+    PrintKernelSuccess("ACPI: All processes and services stopped\n");
 }
 
 void ACPIShutdown(void) {
-    if (!g_acpi_initialized || !g_fadt) {
-        PrintKernel("ACPI: Shutdown not available, using fallback\n");
-        outw(0x604, 0x2000);
-        return;
-    }
-    
     PrintKernel("ACPI: Initiating shutdown...\n");
 
     // Enable ACPI mode if needed
     if (g_fadt->smi_command_port && g_fadt->acpi_enable) {
         PrintKernel("ACPI: Enabling ACPI mode via SMI\n");
         outb(g_fadt->smi_command_port, g_fadt->acpi_enable);
     }
-    
+
+    ACPIResetProcedure();
+
     // Try multiple shutdown methods
-    uint16_t shutdown_values[] = {0x2000, 0x3C00, 0x1400, 0x0000};
-    
     for (int i = 0; i < 4; i++) {
+        const uint16_t shutdown_values[] = {0x2000, 0x3C00, 0x1400, 0x0000};
         PrintKernel("ACPI: Trying shutdown value 0x");
         PrintKernelHex(shutdown_values[i]);
         PrintKernel(" on port 0x");
@@ -180,18 +180,17 @@ void ACPIShutdown(void) {
         // Wait a bit
         delay(10);
     }
-    
-    // Fallback methods
-    PrintKernel("ACPI: Trying QEMU shutdown\n");
-    outw(0x604, 0x2000);
-    
+
     PrintKernel("ACPI: Trying Bochs shutdown\n");
     outw(0xB004, 0x2000);
 
     PrintKernelError("ACPI: All shutdown methods failed\n");
 }
 
 void ACPIReboot(void) {
+
+    ACPIResetProcedure();
+
     PrintKernel("ACPI: Initiating reboot...\n");
 
     // Try keyboard controller reset

drivers/ACPI.h

@@ -82,4 +82,5 @@ typedef struct {
 // Function declarations
 bool ACPIInit(void);
 void ACPIShutdown(void);
-void ACPIReboot(void);
+void ACPIReboot(void);
+void* AcpiFindTable(const char* signature);

fs/EXT/Ext2.c

@@ -118,7 +118,7 @@ int Ext2ReadBlock(uint32_t block, void* buffer) {
     return 0;
 }
 
-static FileSystemDriver ext2_driver = {"EXT2", Ext2Detect, Ext2Mount};
+static FileSystemDriver ext2_driver = {"EXT2", Ext2Detect, Ext2Mount, Ext2Unmount};
 
 int Ext2Mount(BlockDevice* device, const char* mount_point) {
     // Prepare or reuse per-device volume
@@ -222,11 +222,37 @@ int Ext2Mount(BlockDevice* device, const char* mount_point) {
     }
     PrintKernelF("EXT2: Mounted filesystem\n");
 
-    PrintKernelSuccess("EXT2: Filesystem initialized successfully.\n");
+    PrintKernelSuccess("EXT2: Filesystem mounted.\n");
     rust_rwlock_write_unlock(volume.lock);
     return 0;
 }
 
+int Ext2Unmount(BlockDevice* device) {
+    if (!device) { return -1; }
+    int device_id = device->id;
+    if (device_id < 0 || device_id >= MAX_BLOCK_DEVICES) { return -1; }
+    Ext2Volume* vol = g_ext2_by_dev[device_id];
+    if (!vol) { return -1; } // Not mounted
+    RustRwLock* lock = vol->lock;
+    if (lock) {
+        rust_rwlock_write_lock(lock, GetCurrentProcess()->pid);
+    }
+    if (g_ext2_active == vol) {
+        g_ext2_active = NULL;
+    }
+    if (vol->group_descs) {
+        KernelFree(vol->group_descs);
+        vol->group_descs = NULL;
+    }
+    g_ext2_by_dev[device_id] = NULL;
+    if (lock) {
+        rust_rwlock_write_unlock(lock);
+        rust_rwlock_free(lock);
+    }
+    KernelFree(vol);
+    return 0;
+}
+
 int Ext2ReadInode(uint32_t inode_num, Ext2Inode* inode) {
     rust_rwlock_read_lock(volume.lock, GetCurrentProcess()->pid);
     if (inode_num == 0) {

fs/EXT/Ext2.h

@@ -101,6 +101,7 @@ typedef struct {
 
 // Function prototypes for VFS integration
 int Ext2Mount(BlockDevice* device, const char* mount_point);
+int Ext2Unmount(BlockDevice* device);
 int Ext2Detect(BlockDevice* device);
 void Ext2SetActive(BlockDevice* device);
 int Ext2ReadFile(const char* path, void* buffer, uint32_t max_size);