feature (proc): implemented basic process structure

kernel/inc/proc/process.h

@@ -0,0 +1,51 @@
+/*
+ * Copyright 2023 Antifallobst <antifallobst@systemausfall.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy of
+ * this software and associated documentation files (the “Software”), to deal in
+ * the Software without restriction, including without limitation the rights to use,
+ * copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the
+ * Software, and to permit persons to whom the Software is furnished to do so,
+ * subject to the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
+ * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+ * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+ * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef NOX_PROCESS_H
+#define NOX_PROCESS_H
+
+#include "utils/stdtypes.h"
+#include "utils/string.h"
+
+typedef uint32_t pid_t;
+
+typedef enum {
+    PROCESS_NONE,
+    PROCESS_KERNEL
+} processes_standard_E;
+
+typedef struct process_T process_T;
+struct process_T{
+    char        name[128];
+    pid_t       id;
+    void*       chunk;
+    uint32_t    chunk_id;
+
+    process_T*  parent;
+    void*       threads;
+};
+
+pid_t   process_spawn   (pid_t parent, string_t name);
+void    process_kill    (pid_t process);
+
+#endif //NOX_PROCESS_H

kernel/inc/proc/scheduler.h

@@ -25,11 +25,27 @@
 #define NOX_SCHEDULER_H
 
 #include "proc/thread.h"
+#include "proc/process.h"
 #include "utils/bitmap.h"
 
+#define SCHEDULER_PROCESS_CHUNK_SIZE 64
+
+typedef struct scheduler_processes_chunk_T scheduler_processes_chunk_T;
+struct scheduler_processes_chunk_T {
+    process_T**                     processes;
+    bitmap_T                        processes_bitmap;
+
+    uint32_t                        num_free_pids;
+
+    scheduler_processes_chunk_T*    prev;
+    scheduler_processes_chunk_T*    next;
+};
+
 typedef struct {
     uint32_t                        num_threads;
+    uint32_t                        num_processes;
     thread_T*                       running_thread;
+    scheduler_processes_chunk_T*    processes;
 
     bool                            blocked;
     bool                            initialized;
@@ -44,7 +60,12 @@ void            scheduler_pause_thread          (thread_T* thread);
 void            scheduler_start_thread              (thread_T* thread);
 void            scheduler_kill_thread               (thread_T* thread);
 
+pid_t           scheduler_register_process          (process_T* process);
+void            scheduler_kill_process              (process_T* process);
+process_T*      scheduler_get_process               (pid_t pid);
+
 thread_T*       scheduler_get_current_thread        ();
+process_T*      scheduler_get_current_process       ();
 
 cpu_state_T*    scheduler_switch_context            (cpu_state_T* state);
 

kernel/inc/proc/thread.h

@@ -26,6 +26,7 @@
 
 #include "utils/stdtypes.h"
 #include "platform/cpu.h"
+#include "proc/process.h"
 
 typedef struct thread_T thread_T;
 struct thread_T{
@@ -33,15 +34,19 @@ struct thread_T{
     uint64_t                cpu_time;
     void*                   stack;
     uint32_t                stack_size;
+    process_T*              process;
 
     // Scheduling data
-    thread_T*    prev;
+    thread_T*    global_prev;
-    thread_T*    next;
+    thread_T*    global_next;
+
+    thread_T*    local_prev;
+    thread_T*    local_next;
 };
 
 
-thread_T*   thread_spawn            (void* function);
+thread_T*   thread_spawn            (pid_t process, void* function);
-thread_T*   thread_spawn_from_state (cpu_state_T* state);
+thread_T*   thread_spawn_from_state (pid_t process, cpu_state_T* state);
 void        thread_start            (thread_T* thread);
 void        thread_pause            (thread_T* thread);
 void        thread_kill             (thread_T* thread);

kernel/src/kmain.c

@@ -78,6 +78,26 @@ void kernel_init(boot_info_T* boot_info) {
     limine_terminal_print(boot_info, " ok\n");
 }
 
+#include "utils/io.h"
+
+void test1() {
+    while (true) {
+        io_out_byte(LOG_PORT, 'A');
+    }
+}
+
+void test2() {
+    while (true) {
+        io_out_byte(LOG_PORT, 'B');
+    }
+}
+
+void test3() {
+    while (true) {
+        io_out_byte(LOG_PORT, 'C');
+    }
+}
+
 void kmain(boot_info_T boot_info) {
 
     limine_terminal_print(&boot_info, "Booting NoxOS...\n");
@@ -90,16 +110,29 @@ void kmain(boot_info_T boot_info) {
 
     elf_init_kernel_exec(&boot_info);
 
-    vfs_node_T* node = vfs_resolve_path(&g_root_fs, "/initrd/test.elf");
+    pid_t proc1 = process_spawn(PROCESS_NONE, "test process 1");
+    thread_T* proc1_thread1 = thread_spawn(proc1, test1);
+    thread_T* proc1_thread2 = thread_spawn(proc1, test2);
+    thread_T* proc1_thread3 = thread_spawn(proc1, test3);
 
-        elf_executable_T* exec = elf_executable_create(node->cache->buffer);
+    thread_start(proc1_thread1);
+    thread_start(proc1_thread2);
+    thread_start(proc1_thread3);
 
-    page_map_T* page_map = pframe_request();
+    process_kill(proc1);
-    elf_mappings_apply(exec->mappings, exec->num_mappings, node->cache->buffer, 0xFFFFFFFFFF000000, page_map);
 
-    int (*entry)() = (0xFFFFFFFFFF000000 + symbol_resolve_from_name(exec->symbols, exec->num_symbols, "_start")->address);
+//    vfs_node_T* node = vfs_resolve_path(&g_root_fs, "/initrd/test.elf");
+//
+//    elf_executable_T* exec = elf_executable_create(node->cache->buffer);
+//
+//    page_map_T* page_map = pframe_request();
+//    elf_mappings_apply(exec->mappings, exec->num_mappings, node->cache->buffer, 0xFFFFFFFFFF000000, page_map);
+//
+//    int (*entry)() = (0xFFFFFFFFFF000000 + symbol_resolve_from_name(exec->symbols, exec->num_symbols, "_start")->address);
+//
+//    thread_spawn(PROCESS_KERNEL, entry);
 
-    log(LOG_DEBUG, "ELF returned: %d", entry());
+//    log(LOG_DEBUG, "ELF returned: %d", entry());
 
     CORE_HALT_FOREVER
 }

kernel/src/proc/process.c

@@ -0,0 +1,47 @@
+/*
+ * Copyright 2023 Antifallobst <antifallobst@systemausfall.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy of
+ * this software and associated documentation files (the “Software”), to deal in
+ * the Software without restriction, including without limitation the rights to use,
+ * copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the
+ * Software, and to permit persons to whom the Software is furnished to do so,
+ * subject to the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
+ * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+ * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+ * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "proc/process.h"
+#include "proc/scheduler.h"
+#include "utils/memory.h"
+#include "utils/math.h"
+
+pid_t process_spawn(pid_t parent, string_t name) {
+    process_T* process = memory_allocate(sizeof(process_T));
+    process->threads = NULL;
+    process->parent  = scheduler_get_process(parent);
+
+    memory_copy(name, process->name, MIN(string_length(name), 127));
+
+    // the processes' id, chunk_id and chunk fields will be set by the scheduler
+
+    return scheduler_register_process(process);
+}
+
+void process_kill(pid_t process) {
+    process_T* proc = scheduler_get_process(process);
+
+    scheduler_kill_process(proc);
+
+    memory_free(proc);
+}

kernel/src/proc/scheduler.c

@@ -25,22 +25,52 @@
 #include "utils/memory.h"
 #include "utils/logger.h"
 #include "utils/core.h"
+#include "drivers/graphics/renderer.h"
 #include "platform/syscall.h"
 
 scheduler_T         g_scheduler;
+graphics_buffer_T*  g_scheduler_info_buffer;
+
+void scheduler_update_info() {
+    uint64_t size = format_size("Processes[%d]\n  Threads[%d]", g_scheduler.num_processes, g_scheduler.num_threads);
+    char buffer[size];
+    format(buffer, "Processes[%d]\n  Threads[%d]", g_scheduler.num_processes, g_scheduler.num_threads);
+
+    graphics_buffer_draw_string(g_scheduler_info_buffer, 0, 0, g_color_palette[COLOR_PAL_GREEN], buffer);
+}
+
+scheduler_processes_chunk_T* scheduler_processes_chunk_create(scheduler_processes_chunk_T* prev) {
+    scheduler_processes_chunk_T* chunk = memory_allocate(sizeof(scheduler_processes_chunk_T));
+
+    chunk->processes        = memory_allocate(SCHEDULER_PROCESS_CHUNK_SIZE * sizeof(process_T*));
+    chunk->prev             = prev;
+    chunk->next             = NULL;
+    chunk->processes_bitmap = bitmap_init(SCHEDULER_PROCESS_CHUNK_SIZE);
+    chunk->num_free_pids    = SCHEDULER_PROCESS_CHUNK_SIZE;
+
+    return chunk;
+}
 
 void scheduler_init() {
     g_scheduler.num_threads     = 0;
+    g_scheduler.num_processes   = 0;
     g_scheduler.running_thread  = NULL;
     g_scheduler.blocked         = false;
 
+    g_scheduler.processes       = scheduler_processes_chunk_create(NULL);
+
+    g_scheduler_info_buffer = graphics_buffer_request(graphics_renderer_get_width()-100, 0, 100, 200, GRAPHICS_BUFFER_STANDARD);
+
     syscall_perform(SYSCALL_KERNEL_SCHEDULER_START);
 }
 
 cpu_state_T* scheduler_start(cpu_state_T* state) {
-    thread_T* thread = thread_spawn_from_state(state);
+
-    thread->prev = thread;
+    process_spawn(PROCESS_NONE, "kernel");
-    thread->next = thread;
+
+    thread_T* thread = thread_spawn_from_state(PROCESS_KERNEL, state);
+    thread->global_prev = thread;
+    thread->global_next = thread;
     g_scheduler.running_thread = thread;
 
 
@@ -58,25 +88,36 @@ bool scheduler_is_initialized() {
 void scheduler_queue_add_thread(thread_T* thread) {
     if (g_scheduler.running_thread == NULL) { return; }
 
-    thread->prev                           = g_scheduler.running_thread;
+    thread->global_prev                                  = g_scheduler.running_thread;
-    thread->next                           = g_scheduler.running_thread->next;
+    thread->global_next                                  = g_scheduler.running_thread->global_next;
-    g_scheduler.running_thread->next->prev = thread;
+    g_scheduler.running_thread->global_next->global_prev = thread;
-    g_scheduler.running_thread->next       = thread;
+    g_scheduler.running_thread->global_next              = thread;
 }
 
 void scheduler_queue_remove_thread(thread_T* thread) {
-    if (thread->prev == NULL || thread->next == NULL) { return; }
+    if (thread->global_prev == NULL || thread->global_next == NULL) { return; }
 
-    thread->prev->next = thread->next;
+    thread->global_prev->global_next = thread->global_next;
-    thread->next->prev = thread->prev;
+    thread->global_next->global_prev = thread->global_prev;
-    thread->prev       = NULL;
+    thread->global_prev              = NULL;
-    thread->next       = NULL;
+    thread->global_next              = NULL;
 }
 
 thread_T* scheduler_register_thread(thread_T* thread) {
+    thread->local_prev = NULL;
+    thread->local_next = NULL;
+
+    if (thread->process->threads != NULL) {
+        thread->local_next             = thread->process->threads;
+        thread->local_next->local_prev = thread;
+    }
+
+    thread->process->threads = thread;
+
     g_scheduler.num_threads++;
 
     log(LOG_INFO, "<Scheduler> Registered thread");
+    scheduler_update_info();
 
     return thread;
 }
@@ -84,25 +125,129 @@ thread_T* scheduler_register_thread(thread_T* thread) {
 void scheduler_pause_thread(thread_T* thread) {
     scheduler_queue_remove_thread(thread);
     log(LOG_INFO, "<Scheduler> Paused thread");
+    scheduler_update_info();
 }
 
 void scheduler_start_thread(thread_T* thread) {
     scheduler_queue_add_thread(thread);
     log(LOG_INFO, "<Scheduler> Started thread");
+    scheduler_update_info();
 }
 
 void scheduler_kill_thread(thread_T* thread) {
     scheduler_queue_remove_thread(thread);
 
+    if (thread->process->threads == thread) {
+        thread->process->threads = thread->local_next;
+    }
+
+    if (thread->local_next != NULL) {
+        thread->local_next->local_prev = thread->local_prev;
+    }
+
+    if (thread->local_prev != NULL) {
+        thread->local_prev->local_next = thread->local_next;
+    }
+
     g_scheduler.num_threads--;
 
     log(LOG_INFO, "<Scheduler> Killed thread");
+    scheduler_update_info();
+}
+
+pid_t scheduler_register_process(process_T* process) {
+    pid_t pid = 1;  // offset pit by one, so this won't return a pid that is PROCESS_NONE
+
+    // find a process chunk, that has free slots
+    scheduler_processes_chunk_T* chunk = g_scheduler.processes;
+    while (chunk->num_free_pids == 0) {
+        if (chunk->next == NULL) {
+            chunk->next = scheduler_processes_chunk_create(chunk);
+        }
+        chunk  = chunk->next;
+        pid   += SCHEDULER_PROCESS_CHUNK_SIZE;
+    }
+
+    for (uint32_t i = 0; i < SCHEDULER_PROCESS_CHUNK_SIZE; i++) {
+        if (bitmap_get(&chunk->processes_bitmap, i)) {
+            continue;
+        }
+
+        bitmap_set(&chunk->processes_bitmap, i, true);
+        chunk->processes[i] = process;
+        process->chunk      = chunk;
+        process->chunk_id   = i;
+        process->id         = pid + i;
+
+        chunk->num_free_pids--;
+        g_scheduler.num_processes++;
+
+        log(LOG_INFO, "<Scheduler> Registered Process %d", pid + i);
+        scheduler_update_info();
+
+        return pid + i;
+    }
+
+    // if this happens, something went really wrong
+    log(LOG_ERROR, "<Scheduler> failed to register process (no free slots in chunk)");
+    return PROCESS_NONE;
+}
+
+void scheduler_kill_process(process_T* process) {
+    scheduler_processes_chunk_T* chunk = process->chunk;
+
+    bitmap_set(&chunk->processes_bitmap, process->chunk_id, false);
+    chunk->processes[process->chunk_id] = NULL;
+
+
+    thread_T* thread = process->threads;
+    while (thread != NULL) {
+        thread_T* next = thread->local_next;
+        thread_kill(thread);
+        thread = next;
+    }
+
+    chunk->num_free_pids++;
+    g_scheduler.num_processes--;
+
+    log(LOG_INFO, "<Scheduler> Killed Process %d", process->id);
+    scheduler_update_info();
+}
+
+process_T* scheduler_get_process(pid_t pid) {
+    if (pid == PROCESS_NONE) {
+        return NULL;
+    }
+
+    pid--;  // apply pid offset to make PROCESS_NONE work
+
+    uint32_t num_chunks     = pid / SCHEDULER_PROCESS_CHUNK_SIZE;
+    uint32_t index_in_chunk = pid % SCHEDULER_PROCESS_CHUNK_SIZE;
+
+    scheduler_processes_chunk_T* chunk = g_scheduler.processes;
+    for (uint32_t i = 0; i < num_chunks; i++) {
+        chunk = chunk->next;
+
+        if (chunk == NULL) {
+            return NULL;
+        }
+    }
+
+    if (!bitmap_get(&chunk->processes_bitmap, index_in_chunk)) {
+        return NULL;
+    }
+
+    return chunk->processes[index_in_chunk];
 }
 
 thread_T* scheduler_get_current_thread() {
     return g_scheduler.running_thread;
 }
 
+process_T* scheduler_get_current_process() {
+    return g_scheduler.running_thread->process;
+}
+
 cpu_state_T* scheduler_switch_context(cpu_state_T* state) {
     if (!g_scheduler.initialized) {
         return state;
@@ -112,14 +257,14 @@ cpu_state_T* scheduler_switch_context(cpu_state_T* state) {
     g_scheduler.blocked = true;
 
     thread_T* old_thread = g_scheduler.running_thread;
-    thread_T* new_thread = g_scheduler.running_thread->next;
+    thread_T* new_thread = g_scheduler.running_thread->global_next;
 
     if (old_thread->cpu_time > 0) {
         memory_copy(state, &old_thread->state, sizeof(cpu_state_T));
     }
 
     old_thread->cpu_time += 1;
-    g_scheduler.running_thread = g_scheduler.running_thread->next;
+    g_scheduler.running_thread = g_scheduler.running_thread->global_next;
 
     g_scheduler.blocked = false;
     return &new_thread->state;

kernel/src/proc/thread.c

@@ -27,10 +27,17 @@
 #include "mm/page_frame.h"
 #include "platform/gdt.h"
 #include "utils/memory.h"
+#include "utils/logger.h"
 
-thread_T* thread_spawn(void* function) {
+thread_T* thread_spawn(pid_t process, void* function) {
     thread_T* thread = memory_allocate(sizeof(thread_T));
 
+    thread->process = scheduler_get_process(process);
+    if (thread->process == NULL) {
+        log(LOG_WARNING, "<Thread> failed to spawn thread for process %d (invalid pid)", process);
+        return NULL;
+    }
+
     thread->stack_size = PFRAME_SIZE * 4;
     thread->stack      = memory_allocate(thread->stack_size);
     thread->cpu_time   = 0;
@@ -56,9 +63,15 @@ thread_T* thread_spawn(void* function) {
     return scheduler_register_thread(thread);
 }
 
-thread_T* thread_spawn_from_state(cpu_state_T* state) {
+thread_T* thread_spawn_from_state(pid_t process, cpu_state_T* state) {
     thread_T* thread = memory_allocate(sizeof(thread_T));
 
+    thread->process = scheduler_get_process(process);
+    if (thread->process == NULL) {
+        log(LOG_WARNING, "<Thread> failed to spawn thread for process %d (invalid pid)", process);
+        return NULL;
+    }
+
     thread->stack_size = 0;
     thread->stack      = NULL;
     thread->cpu_time   = 0;