CHECKP: shitty VFS

src/fs/fat32.h

@@ -25,7 +25,7 @@ typedef struct {
     uint16_t num_heads;             // Number of heads
     uint32_t hidden_sectors;        // Hidden sectors
     uint32_t total_sectors_32;      // Total sectors 32-bit
-    
+
     // FAT32 Specific
     uint32_t sectors_per_fat_32;    // Sectors per FAT 32-bit
     uint16_t flags;                 // Flags
@@ -61,6 +61,18 @@ typedef struct {
     uint32_t file_size;             // File size
 } __attribute__((packed)) FAT32_DirEntry;
 
+// Long File Name Directory Entry (32 bytes)
+typedef struct {
+    uint8_t order;                  // Sequence number (0x40 = last, | index)
+    uint16_t name1[5];              // Characters 1-5 (UCS-2)
+    uint8_t attr;                   // Always 0x0F
+    uint8_t type;                   // Always 0x00
+    uint8_t checksum;               // Checksum of short name
+    uint16_t name2[6];              // Characters 6-11 (UCS-2)
+    uint16_t first_cluster;         // Always 0x0000
+    uint16_t name3[2];              // Characters 12-13 (UCS-2)
+} __attribute__((packed)) FAT32_LFNEntry;
+
 // File Attributes
 #define ATTR_READ_ONLY   0x01
 #define ATTR_HIDDEN      0x02
@@ -70,6 +82,7 @@ typedef struct {
 #define ATTR_ARCHIVE     0x20
 #define ATTR_DEVICE      0x40
 #define ATTR_RESERVED    0x80
+#define ATTR_LFN         0x0F  // LFN marker (all of the above ORed)
 
 // FAT32 Constants
 #define FAT32_SECTOR_SIZE 512
@@ -88,7 +101,7 @@ typedef struct {
     bool valid;                     // Is this handle valid?
     uint32_t dir_sector;            // Sector containing the directory entry
     uint32_t dir_offset;            // Offset within that sector
-    char drive;                     // Drive letter (A, B, ...)
+    void *volume;                   // Pointer to owning FAT32_Volume (or NULL for ramfs)
 } FAT32_FileHandle;
 
 // Directory Entry Info (for listing)
@@ -101,12 +114,23 @@ typedef struct {
     uint16_t write_time;
 } FAT32_FileInfo;
 
+// === VFS Integration ===
+// Forward-declared VFS ops type (defined in vfs.h)
+struct vfs_fs_ops;
+
+// Get VFS ops structs for registration
+struct vfs_fs_ops* fat32_get_ramfs_ops(void);
+struct vfs_fs_ops* fat32_get_realfs_ops(void);
+
+// Mount a real FAT32 volume from a block device — returns fs_private for VFS
+void* fat32_mount_volume(void *disk_ptr);
+
 // === Function Declarations ===
 
 // Initialization
 void fat32_init(void);
 
-// File Operations
+// File Operations (backward-compat wrappers — dispatch through VFS)
 FAT32_FileHandle* fat32_open(const char *path, const char *mode);
 void fat32_close(FAT32_FileHandle *handle);
 int fat32_read(FAT32_FileHandle *handle, void *buffer, int size);
@@ -124,7 +148,10 @@ bool fat32_is_directory(const char *path);
 // Listing
 int fat32_list_directory(const char *path, FAT32_FileInfo *entries, int max_entries);
 
-// Working Directory
+// Info
+int fat32_get_info(const char *path, FAT32_FileInfo *info);
+
+// Working Directory (backward compat — wraps VFS path tracking)
 bool fat32_chdir(const char *path);
 void fat32_get_current_dir(char *buffer, int size);
 bool fat32_change_drive(char drive);

src/fs/vfs.c

@@ -0,0 +1,760 @@
+// Copyright (c) 2023-2026 Chris (boreddevnl)
+// This software is released under the GNU General Public License v3.0. See LICENSE file for details.
+// This header needs to maintain in any file it is present in, as per the GPL license terms.
+#include "vfs.h"
+#include "memory_manager.h"
+#include "spinlock.h"
+#include <stddef.h>
+#include "disk.h"
+
+// ============================================================================
+// VFS Mount Table and File Handle Pool
+// ============================================================================
+
+static vfs_mount_t mounts[VFS_MAX_MOUNTS];
+static int mount_count = 0;
+static vfs_file_t open_files[VFS_MAX_OPEN_FILES];
+static spinlock_t vfs_lock = SPINLOCK_INIT;
+
+extern void serial_write(const char *str);
+extern void serial_write_num(uint64_t num);
+
+// ============================================================================
+// String helpers (freestanding — no libc)
+// ============================================================================
+
+static int vfs_strlen(const char *s) {
+    int n = 0;
+    while (s[n]) n++;
+    return n;
+}
+
+static void vfs_strcpy(char *d, const char *s) {
+    while ((*d++ = *s++));
+}
+
+static int vfs_strcmp(const char *a, const char *b) {
+    while (*a && *a == *b) { a++; b++; }
+    return (unsigned char)*a - (unsigned char)*b;
+}
+
+static int vfs_strncmp(const char *a, const char *b, int n) {
+    for (int i = 0; i < n; i++) {
+        if (a[i] != b[i]) return (unsigned char)a[i] - (unsigned char)b[i];
+        if (!a[i]) return 0;
+    }
+    return 0;
+}
+
+static bool vfs_starts_with(const char *str, const char *prefix) {
+    while (*prefix) {
+        if (*str++ != *prefix++) return false;
+    }
+    return true;
+}
+
+// ============================================================================
+// Path Normalization
+// ============================================================================
+
+void vfs_normalize_path(const char *path, char *normalized) {
+    // Resolve . and .. components, remove duplicate slashes
+    char parts[32][128];
+    int depth = 0;
+    int i = 0;
+
+    // Skip leading slash
+    if (path[0] == '/') i = 1;
+
+    while (path[i]) {
+        // Skip duplicate slashes
+        if (path[i] == '/') { i++; continue; }
+
+        // Extract component
+        int j = 0;
+        while (path[i] && path[i] != '/' && j < 127) {
+            parts[depth][j++] = path[i++];
+        }
+        parts[depth][j] = 0;
+
+        if (parts[depth][0] == '.' && parts[depth][1] == 0) {
+            // "." — skip
+        } else if (parts[depth][0] == '.' && parts[depth][1] == '.' && parts[depth][2] == 0) {
+            // ".." — go up
+            if (depth > 0) depth--;
+        } else {
+            depth++;
+            if (depth >= 32) break;
+        }
+
+        if (path[i] == '/') i++;
+    }
+
+    // Reconstruct
+    normalized[0] = '/';
+    int pos = 1;
+    for (int k = 0; k < depth; k++) {
+        int l = 0;
+        while (parts[k][l] && pos < VFS_MAX_PATH - 2) {
+            normalized[pos++] = parts[k][l++];
+        }
+        if (k < depth - 1) normalized[pos++] = '/';
+    }
+    normalized[pos] = 0;
+
+    // Ensure root is just "/"
+    if (pos == 1 && normalized[0] == '/') {
+        normalized[1] = 0;
+    }
+}
+
+// ============================================================================
+// Mount Resolution — find the longest-prefix mount for a path
+// ============================================================================
+
+static vfs_mount_t* vfs_resolve_mount(const char *path, const char **rel_path_out) {
+    vfs_mount_t *best = NULL;
+    int best_len = -1;
+
+    for (int i = 0; i < mount_count; i++) {
+        if (!mounts[i].active) continue;
+
+        int mlen = mounts[i].path_len;
+
+        // Root mount "/" matches everything
+        if (mlen == 1 && mounts[i].path[0] == '/') {
+            if (best_len < 1) {
+                best = &mounts[i];
+                best_len = 1;
+            }
+            continue;
+        }
+
+        // Check if path starts with this mount point
+        if (vfs_strncmp(path, mounts[i].path, mlen) == 0) {
+            // Must be followed by '/' or end of string
+            if (path[mlen] == '/' || path[mlen] == '\0') {
+                if (mlen > best_len) {
+                    best = &mounts[i];
+                    best_len = mlen;
+                }
+            }
+        }
+    }
+
+    if (best && rel_path_out) {
+        const char *rel = path + best_len;
+        // Skip leading slash in relative path
+        while (*rel == '/') rel++;
+        *rel_path_out = rel;
+    }
+
+    return best;
+}
+
+// ============================================================================
+// File Handle Pool
+// ============================================================================
+
+static vfs_file_t* vfs_alloc_file(void) {
+    for (int i = 0; i < VFS_MAX_OPEN_FILES; i++) {
+        if (!open_files[i].valid) {
+            open_files[i].valid = true;
+            return &open_files[i];
+        }
+    }
+    return NULL;
+}
+
+static void vfs_free_file(vfs_file_t *f) {
+    if (f) {
+        f->valid = false;
+        f->fs_handle = NULL;
+        f->mount = NULL;
+    }
+}
+
+// ============================================================================
+// Initialization
+// ============================================================================
+
+void vfs_init(void) {
+    for (int i = 0; i < VFS_MAX_MOUNTS; i++) {
+        mounts[i].active = false;
+    }
+    for (int i = 0; i < VFS_MAX_OPEN_FILES; i++) {
+        open_files[i].valid = false;
+    }
+    mount_count = 0;
+
+    serial_write("[VFS] Virtual File System initialized\n");
+}
+
+// ============================================================================
+// Mount / Unmount
+// ============================================================================
+
+bool vfs_mount(const char *mount_path, const char *device, const char *fs_type,
+               vfs_fs_ops_t *ops, void *fs_private) {
+    uint64_t flags = spinlock_acquire_irqsave(&vfs_lock);
+
+    if (mount_count >= VFS_MAX_MOUNTS) {
+        spinlock_release_irqrestore(&vfs_lock, flags);
+        serial_write("[VFS] ERROR: Mount table full\n");
+        return false;
+    }
+
+    // Check for duplicate mount
+    for (int i = 0; i < mount_count; i++) {
+        if (mounts[i].active && vfs_strcmp(mounts[i].path, mount_path) == 0) {
+            spinlock_release_irqrestore(&vfs_lock, flags);
+            serial_write("[VFS] ERROR: Mount point already in use: ");
+            serial_write(mount_path);
+            serial_write("\n");
+            return false;
+        }
+    }
+
+    vfs_mount_t *m = &mounts[mount_count];
+    vfs_strcpy(m->path, mount_path);
+    m->path_len = vfs_strlen(mount_path);
+    m->ops = ops;
+    m->fs_private = fs_private;
+    vfs_strcpy(m->device, device ? device : "none");
+    vfs_strcpy(m->fs_type, fs_type ? fs_type : "unknown");
+    m->active = true;
+    mount_count++;
+
+    spinlock_release_irqrestore(&vfs_lock, flags);
+
+    serial_write("[VFS] Mounted ");
+    serial_write(fs_type);
+    serial_write(" (");
+    serial_write(device ? device : "none");
+    serial_write(") at ");
+    serial_write(mount_path);
+    serial_write("\n");
+
+    return true;
+}
+
+bool vfs_umount(const char *mount_path) {
+    uint64_t flags = spinlock_acquire_irqsave(&vfs_lock);
+
+    for (int i = 0; i < mount_count; i++) {
+        if (mounts[i].active && vfs_strcmp(mounts[i].path, mount_path) == 0) {
+            // Close any open files on this mount
+            for (int j = 0; j < VFS_MAX_OPEN_FILES; j++) {
+                if (open_files[j].valid && open_files[j].mount == &mounts[i]) {
+                    if (mounts[i].ops->close) {
+                        mounts[i].ops->close(mounts[i].fs_private, open_files[j].fs_handle);
+                    }
+                    vfs_free_file(&open_files[j]);
+                }
+            }
+
+            serial_write("[VFS] Unmounted ");
+            serial_write(mounts[i].path);
+            serial_write("\n");
+
+            mounts[i].active = false;
+
+            // Compact array
+            for (int k = i; k < mount_count - 1; k++) {
+                mounts[k] = mounts[k + 1];
+            }
+            mount_count--;
+
+            spinlock_release_irqrestore(&vfs_lock, flags);
+            return true;
+        }
+    }
+
+    spinlock_release_irqrestore(&vfs_lock, flags);
+    return false;
+}
+
+// ============================================================================
+// File Operations
+// ============================================================================
+
+vfs_file_t* vfs_open(const char *path, const char *mode) {
+    if (!path || !mode) return NULL;
+
+    // Normalize path
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    uint64_t flags = spinlock_acquire_irqsave(&vfs_lock);
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    if (!mount || !mount->ops->open) {
+        spinlock_release_irqrestore(&vfs_lock, flags);
+        return NULL;
+    }
+
+    // If rel_path is empty, use root
+    if (!rel_path || rel_path[0] == '\0') {
+        rel_path = "/";
+    }
+
+    vfs_file_t *vf = vfs_alloc_file();
+    if (!vf) {
+        spinlock_release_irqrestore(&vfs_lock, flags);
+        serial_write("[VFS] ERROR: No free file handles\n");
+        return NULL;
+    }
+
+    vf->mount = mount;
+
+    // Release lock before calling FS ops (FS has its own locking)
+    spinlock_release_irqrestore(&vfs_lock, flags);
+
+    void *fs_handle = mount->ops->open(mount->fs_private, rel_path, mode);
+    if (!fs_handle) {
+        flags = spinlock_acquire_irqsave(&vfs_lock);
+        vfs_free_file(vf);
+        spinlock_release_irqrestore(&vfs_lock, flags);
+        return NULL;
+    }
+
+    vf->fs_handle = fs_handle;
+    return vf;
+}
+
+void vfs_close(vfs_file_t *file) {
+    if (!file || !file->valid) return;
+
+    vfs_mount_t *mount = file->mount;
+    if (mount && mount->ops->close) {
+        mount->ops->close(mount->fs_private, file->fs_handle);
+    }
+
+    uint64_t flags = spinlock_acquire_irqsave(&vfs_lock);
+    vfs_free_file(file);
+    spinlock_release_irqrestore(&vfs_lock, flags);
+}
+
+int vfs_read(vfs_file_t *file, void *buf, int size) {
+    if (!file || !file->valid || !file->mount) return -1;
+    if (!file->mount->ops->read) return -1;
+    return file->mount->ops->read(file->mount->fs_private, file->fs_handle, buf, size);
+}
+
+int vfs_write(vfs_file_t *file, const void *buf, int size) {
+    if (!file || !file->valid || !file->mount) return -1;
+    if (!file->mount->ops->write) return -1;
+    return file->mount->ops->write(file->mount->fs_private, file->fs_handle, buf, size);
+}
+
+int vfs_seek(vfs_file_t *file, int offset, int whence) {
+    if (!file || !file->valid || !file->mount) return -1;
+    if (!file->mount->ops->seek) return -1;
+    return file->mount->ops->seek(file->mount->fs_private, file->fs_handle, offset, whence);
+}
+
+uint32_t vfs_file_position(vfs_file_t *file) {
+    if (!file || !file->valid || !file->mount) return 0;
+    if (!file->mount->ops->get_position) return 0;
+    return file->mount->ops->get_position(file->fs_handle);
+}
+
+uint32_t vfs_file_size(vfs_file_t *file) {
+    if (!file || !file->valid || !file->mount) return 0;
+    if (!file->mount->ops->get_size) return 0;
+    return file->mount->ops->get_size(file->fs_handle);
+}
+
+// ============================================================================
+// Directory Operations
+// ============================================================================
+
+int vfs_list_directory(const char *path, vfs_dirent_t *entries, int max) {
+    if (!path || !entries) return -1;
+
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    
+    int count = 0;
+    if (mount && mount->ops->readdir) {
+        if (!rel_path || rel_path[0] == '\0') rel_path = "/";
+        count = mount->ops->readdir(mount->fs_private, rel_path, entries, max);
+        if (count < 0) count = 0; // Treat as virtual if readdir fails
+    }
+
+    // Merge in other mount points that are direct children of this path
+    uint64_t v_flags = spinlock_acquire_irqsave(&vfs_lock);
+    for (int i = 0; i < mount_count; i++) {
+        if (!mounts[i].active) continue;
+        if (vfs_strcmp(mounts[i].path, normalized) == 0) continue; // Skip ourselves
+
+        // Check if mount path starts with current path
+        if (vfs_starts_with(mounts[i].path, normalized)) {
+            const char *sub = mounts[i].path + vfs_strlen(normalized);
+            if (*sub == '/') sub++; // skip slash
+
+            if (*sub != '\0') {
+                // Extract first component (direct child name)
+                char comp[VFS_MAX_NAME];
+                int j = 0;
+                while (sub[j] && sub[j] != '/' && j < VFS_MAX_NAME - 1) {
+                    comp[j] = sub[j];
+                    j++;
+                }
+                comp[j] = 0;
+
+                // Check if already in results
+                bool found = false;
+                for (int k = 0; k < count; k++) {
+                    if (vfs_strcmp(entries[k].name, comp) == 0) {
+                        found = true;
+                        break;
+                    }
+                }
+
+                if (!found && count < max) {
+                    vfs_strcpy(entries[count].name, comp);
+                    entries[count].is_directory = 1;
+                    entries[count].size = 0;
+                    entries[count].start_cluster = 0;
+                    count++;
+                }
+            }
+        }
+    }
+    spinlock_release_irqrestore(&vfs_lock, v_flags);
+
+    // Special case: Ensure "dev" is visible in "/"
+    if (vfs_strcmp(normalized, "/") == 0) {
+        bool found_dev = false;
+        for (int i = 0; i < count; i++) {
+            if (vfs_strcmp(entries[i].name, "dev") == 0) {
+                found_dev = true;
+                break;
+            }
+        }
+        if (!found_dev && count < max) {
+            vfs_strcpy(entries[count].name, "dev");
+            entries[count].is_directory = 1;
+            entries[count].size = 0;
+            entries[count].start_cluster = 0;
+            count++;
+        }
+    }
+
+    // Special case: /dev listing for block devices
+    if (vfs_strcmp(normalized, "/dev") == 0) {
+        int dcount = disk_get_count();
+        for (int i = 0; i < dcount && count < max; i++) {
+            Disk *d = disk_get_by_index(i);
+            if (d) {
+                bool found = false;
+                for (int k = 0; k < count; k++) {
+                    if (vfs_strcmp(entries[k].name, d->devname) == 0) {
+                        found = true;
+                        break;
+                    }
+                }
+                if (!found) {
+                    vfs_strcpy(entries[count].name, d->devname);
+                    entries[count].size = d->total_sectors * 512;
+                    entries[count].is_directory = d->is_partition ? 1 : 0;
+                    entries[count].start_cluster = 0;
+                    entries[count].write_date = 0;
+                    entries[count].write_time = 0;
+                    count++;
+                }
+            }
+        }
+    }
+
+    return count;
+}
+
+bool vfs_mkdir(const char *path) {
+    if (!path) return false;
+
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    if (!mount || !mount->ops->mkdir) return false;
+
+    if (!rel_path || rel_path[0] == '\0') return false;
+
+    return mount->ops->mkdir(mount->fs_private, rel_path);
+}
+
+bool vfs_rmdir(const char *path) {
+    if (!path) return false;
+
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    // Protect root and virtual directories
+    if (normalized[0] == '/' && normalized[1] == '\0') return false;
+    if (vfs_strcmp(normalized, "/dev") == 0) return false;
+    if (vfs_starts_with(normalized, "/dev/")) return false;
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    if (!mount || !mount->ops->rmdir) return false;
+
+    if (!rel_path || rel_path[0] == '\0') return false;
+
+    return mount->ops->rmdir(mount->fs_private, rel_path);
+}
+
+bool vfs_delete(const char *path) {
+    if (!path) return false;
+
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    // Protect root and virtual directories
+    if (normalized[0] == '/' && normalized[1] == '\0') return false;
+    if (vfs_strcmp(normalized, "/dev") == 0) return false;
+    if (vfs_starts_with(normalized, "/dev/")) return false;
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    if (!mount || !mount->ops->unlink) return false;
+
+    if (!rel_path || rel_path[0] == '\0') return false;
+
+    return mount->ops->unlink(mount->fs_private, rel_path);
+}
+
+bool vfs_rename(const char *old_path, const char *new_path) {
+    if (!old_path || !new_path) return false;
+
+    char norm_old[VFS_MAX_PATH], norm_new[VFS_MAX_PATH];
+    vfs_normalize_path(old_path, norm_old);
+    vfs_normalize_path(new_path, norm_new);
+
+    const char *rel_old = NULL, *rel_new = NULL;
+    vfs_mount_t *mount_old = vfs_resolve_mount(norm_old, &rel_old);
+    vfs_mount_t *mount_new = vfs_resolve_mount(norm_new, &rel_new);
+
+    // Can only rename within the same mount
+    if (!mount_old || mount_old != mount_new) return false;
+    if (!mount_old->ops->rename) return false;
+
+    if (!rel_old || rel_old[0] == '\0') return false;
+    if (!rel_new || rel_new[0] == '\0') return false;
+
+    return mount_old->ops->rename(mount_old->fs_private, rel_old, rel_new);
+}
+
+// ============================================================================
+// Query Operations
+// ============================================================================
+
+bool vfs_exists(const char *path) {
+    if (!path) return false;
+
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    // Root always exists
+    if (normalized[0] == '/' && normalized[1] == '\0') return true;
+
+    // Check if it's a prefix of any active mount point
+    uint64_t flags_vfs = spinlock_acquire_irqsave(&vfs_lock);
+    for (int i = 0; i < mount_count; i++) {
+        if (mounts[i].active && vfs_starts_with(mounts[i].path, normalized)) {
+            spinlock_release_irqrestore(&vfs_lock, flags_vfs);
+            return true;
+        }
+    }
+    spinlock_release_irqrestore(&vfs_lock, flags_vfs);
+
+    // /dev always exists as a virtual directory
+    if (vfs_strcmp(normalized, "/dev") == 0) return true;
+
+    // Check if it's a device in /dev
+    if (vfs_starts_with(normalized, "/dev/")) {
+        const char *dev = normalized + 5;
+        if (disk_get_by_name(dev)) return true;
+    }
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    if (!mount || !mount->ops->exists) return false;
+
+    if (!rel_path || rel_path[0] == '\0') return true; // Mount point itself exists
+
+    return mount->ops->exists(mount->fs_private, rel_path);
+}
+
+bool vfs_is_directory(const char *path) {
+    if (!path) return false;
+
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    // Root is always a directory
+    if (normalized[0] == '/' && normalized[1] == '\0') return true;
+
+    // Check if it's a prefix of any active mount point (virtual directory)
+    uint64_t flags_vfs = spinlock_acquire_irqsave(&vfs_lock);
+    for (int i = 0; i < mount_count; i++) {
+        if (mounts[i].active && vfs_starts_with(mounts[i].path, normalized)) {
+            // If it matches exactly a mount point, we still need to check if that FS is a dir
+            if (vfs_strcmp(mounts[i].path, normalized) == 0) {
+                // Exact mount point - it is a directory (mount root)
+                spinlock_release_irqrestore(&vfs_lock, flags_vfs);
+                return true;
+            }
+            // Prefix only - it is a virtual intermediate directory
+            spinlock_release_irqrestore(&vfs_lock, flags_vfs);
+            return true;
+        }
+    }
+    spinlock_release_irqrestore(&vfs_lock, flags_vfs);
+
+    // /dev is always a virtual directory
+    if (vfs_strcmp(normalized, "/dev") == 0) return true;
+
+    // Device check
+    if (vfs_starts_with(normalized, "/dev/")) {
+        const char *dev = normalized + 5;
+        Disk *d = disk_get_by_name(dev);
+        if (d) return d->is_partition ? true : false;
+    }
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    if (!mount) return false;
+
+    // If it's a mount point and we're at its root, it definitely exists as a dir
+    if (!rel_path || rel_path[0] == '\0') return true;
+
+    if (!mount->ops->is_dir) return false;
+    return mount->ops->is_dir(mount->fs_private, rel_path);
+}
+
+int vfs_get_info(const char *path, vfs_dirent_t *info) {
+    if (!path || !info) return -1;
+
+    char normalized[VFS_MAX_PATH];
+    vfs_normalize_path(path, normalized);
+
+    // Root info
+    if (normalized[0] == '/' && normalized[1] == '\0') {
+        vfs_strcpy(info->name, "/");
+        info->size = 0;
+        info->is_directory = 1;
+        info->start_cluster = 0;
+        info->write_date = 0;
+        info->write_time = 0;
+        return 0;
+    }
+
+    // /dev virtual directory info
+    if (vfs_strcmp(normalized, "/dev") == 0) {
+        vfs_strcpy(info->name, "dev");
+        info->size = 0;
+        info->is_directory = 1;
+        info->start_cluster = 0;
+        info->write_date = 0;
+        info->write_time = 0;
+        return 0;
+    }
+
+    // Check if it's a prefix of any active mount point (virtual directory)
+    uint64_t flags_vfs = spinlock_acquire_irqsave(&vfs_lock);
+    for (int i = 0; i < mount_count; i++) {
+        if (mounts[i].active && vfs_starts_with(mounts[i].path, normalized)) {
+            if (vfs_strcmp(mounts[i].path, normalized) != 0) {
+                // Virtual intermediate directory
+                // Get component name
+                const char *p = normalized + vfs_strlen(normalized);
+                while (p > normalized && *(p-1) != '/') p--;
+                vfs_strcpy(info->name, p);
+                info->size = 0;
+                info->is_directory = 1;
+                info->start_cluster = 0;
+                info->write_date = 0;
+                info->write_time = 0;
+                spinlock_release_irqrestore(&vfs_lock, flags_vfs);
+                return 0;
+            }
+        }
+    }
+    spinlock_release_irqrestore(&vfs_lock, flags_vfs);
+
+    // Device check
+    if (vfs_starts_with(normalized, "/dev/")) {
+        const char *dev = normalized + 5;
+        Disk *d = disk_get_by_name(dev);
+        if (d) {
+            vfs_strcpy(info->name, d->devname);
+            info->size = d->total_sectors * 512;
+            info->is_directory = d->is_partition ? 1 : 0;
+            info->start_cluster = 0;
+            info->write_date = 0;
+            info->write_time = 0;
+            return 0;
+        }
+    }
+
+    const char *rel_path = NULL;
+    vfs_mount_t *mount = vfs_resolve_mount(normalized, &rel_path);
+    if (!mount || !mount->ops->get_info) return -1;
+
+    if (!rel_path || rel_path[0] == '\0') {
+        // Info about mount root
+        vfs_strcpy(info->name, mount->device);
+        info->size = 0;
+        info->is_directory = 1;
+        info->start_cluster = 0;
+        info->write_date = 0;
+        info->write_time = 0;
+        return 0;
+    }
+
+    return mount->ops->get_info(mount->fs_private, rel_path, info);
+}
+
+// ============================================================================
+// Mount Enumeration
+// ============================================================================
+
+int vfs_get_mount_count(void) {
+    return mount_count;
+}
+
+vfs_mount_t* vfs_get_mount(int index) {
+    if (index < 0 || index >= mount_count) return NULL;
+    if (!mounts[index].active) return NULL;
+    return &mounts[index];
+}
+
+// ============================================================================
+// Auto-Mount (called when a new partition is discovered)
+// ============================================================================
+
+void vfs_automount_partition(const char *devname) {
+    // Build mount point: /mnt/<devname>
+    char mount_path[64] = "/mnt/";
+    int i = 5;
+    const char *d = devname;
+    while (*d && i < 62) mount_path[i++] = *d++;
+    mount_path[i] = 0;
+
+    serial_write("[VFS] Auto-mount requested for ");
+    serial_write(devname);
+    serial_write(" at ");
+    serial_write(mount_path);
+    serial_write("\n");
+
+    // The actual FAT32 volume creation and mount happens in disk_manager
+    // after probing the partition. This function is called by disk_manager
+    // after it has created the FAT32_Volume.
+}

src/fs/vfs.h

@@ -0,0 +1,115 @@
+// Copyright (c) 2023-2026 Chris (boreddevnl)
+// This software is released under the GNU General Public License v3.0. See LICENSE file for details.
+// This header needs to maintain in any file it is present in, as per the GPL license terms.
+#ifndef VFS_H
+#define VFS_H
+
+#include <stdint.h>
+#include <stdbool.h>
+#include <stddef.h>
+
+#define VFS_MAX_PATH 1024
+#define VFS_MAX_NAME 256
+#define VFS_MAX_MOUNTS 16
+#define VFS_MAX_OPEN_FILES 64
+
+// Forward declarations
+typedef struct vfs_mount vfs_mount_t;
+typedef struct vfs_file vfs_file_t;
+
+// Directory entry for readdir
+typedef struct vfs_dirent {
+    char name[VFS_MAX_NAME];
+    uint32_t size;
+    uint8_t is_directory;
+    uint32_t start_cluster;
+    uint16_t write_date;
+    uint16_t write_time;
+} vfs_dirent_t;
+
+// Filesystem operations — implemented by each filesystem type
+typedef struct vfs_fs_ops {
+    // File operations — return opaque FS handle
+    void* (*open)(void *fs_private, const char *rel_path, const char *mode);
+    void  (*close)(void *fs_private, void *file_handle);
+    int   (*read)(void *fs_private, void *file_handle, void *buf, int size);
+    int   (*write)(void *fs_private, void *file_handle, const void *buf, int size);
+    int   (*seek)(void *fs_private, void *file_handle, int offset, int whence);
+
+    // Directory operations
+    int   (*readdir)(void *fs_private, const char *rel_path, vfs_dirent_t *entries, int max);
+    bool  (*mkdir)(void *fs_private, const char *rel_path);
+    bool  (*rmdir)(void *fs_private, const char *rel_path);
+    bool  (*unlink)(void *fs_private, const char *rel_path);
+    bool  (*rename)(void *fs_private, const char *old_path, const char *new_path);
+
+    // Query operations
+    bool  (*exists)(void *fs_private, const char *rel_path);
+    bool  (*is_dir)(void *fs_private, const char *rel_path);
+    int   (*get_info)(void *fs_private, const char *rel_path, vfs_dirent_t *info);
+
+    // Handle info (for backward compat with syscall position/size queries)
+    uint32_t (*get_position)(void *file_handle);
+    uint32_t (*get_size)(void *file_handle);
+} vfs_fs_ops_t;
+
+// VFS file handle
+struct vfs_file {
+    void *fs_handle;        // FS-specific handle (e.g. FAT32_FileHandle*)
+    vfs_mount_t *mount;     // Mount this file belongs to
+    bool valid;
+};
+
+// Mount entry
+struct vfs_mount {
+    char path[256];         // Mount point (e.g. "/", "/mnt/sda1")
+    int path_len;
+    vfs_fs_ops_t *ops;
+    void *fs_private;       // FS-specific data (e.g. FAT32_Volume*)
+    char device[32];        // Device name (e.g. "ramfs", "sda1")
+    char fs_type[16];       // "ramfs", "fat32"
+    bool active;
+};
+
+// Initialization
+void vfs_init(void);
+
+// Mount/unmount
+bool vfs_mount(const char *mount_path, const char *device, const char *fs_type,
+               vfs_fs_ops_t *ops, void *fs_private);
+bool vfs_umount(const char *mount_path);
+
+// File operations
+vfs_file_t* vfs_open(const char *path, const char *mode);
+void vfs_close(vfs_file_t *file);
+int vfs_read(vfs_file_t *file, void *buf, int size);
+int vfs_write(vfs_file_t *file, const void *buf, int size);
+int vfs_seek(vfs_file_t *file, int offset, int whence);
+
+// Directory operations
+int vfs_list_directory(const char *path, vfs_dirent_t *entries, int max);
+bool vfs_mkdir(const char *path);
+bool vfs_rmdir(const char *path);
+bool vfs_delete(const char *path);
+bool vfs_rename(const char *old_path, const char *new_path);
+
+// Query operations
+bool vfs_exists(const char *path);
+bool vfs_is_directory(const char *path);
+int vfs_get_info(const char *path, vfs_dirent_t *info);
+
+// Mount enumeration
+int vfs_get_mount_count(void);
+vfs_mount_t* vfs_get_mount(int index);
+
+// Block device auto-mount
+void vfs_automount_partition(const char *devname);
+
+// Path utilities
+void vfs_normalize_path(const char *path, char *normalized);
+
+// Backward compat: get position/size from vfs_file
+uint32_t vfs_file_position(vfs_file_t *file);
+uint32_t vfs_file_size(vfs_file_t *file);
+
+#endif