WebSVN – HelenOS – Diff – /branches/dd/kernel/generic/src/main/main.c

 /*
  * Copyright (c) 2001-2004 Jakub Jermar
  * All rights reserved.
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
+ *
  * - Redistributions of source code must retain the above copyright
  *   notice, this list of conditions and the following disclaimer.
  * - Redistributions in binary form must reproduce the above copyright
  *   notice, this list of conditions and the following disclaimer in the
  *   documentation and/or other materials provided with the distribution.
  * - The name of the author may not be used to endorse or promote products
  *   derived from this software without specific prior written permission.
+ *
  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 /** @addtogroup main
  * @{
  */
 /**
  * @file
  * @brief   Main initialization kernel function for all processors.
+ *
  * During kernel boot, all processors, after architecture dependent
  * initialization, start executing code found in this file. After
  * bringing up all subsystems, control is passed to scheduler().
+ *
  * The bootstrap processor starts executing main_bsp() while
  * the application processors start executing main_ap().
+ *
  * @see scheduler()
  * @see main_bsp()
  * @see main_ap()
  */
 #include <arch/asm.h>
 #include <context.h>
 #include <print.h>
 #include <panic.h>
 #include <debug.h>
 #include <config.h>
 #include <time/clock.h>
 #include <time/timeout.h>
 #include <proc/scheduler.h>
 #include <proc/thread.h>
 #include <proc/task.h>
 #include <proc/tasklet.h>
 #include <main/kinit.h>
 #include <main/version.h>
 #include <console/kconsole.h>
+#include <console/console.h>
 #include <cpu.h>
 #include <align.h>
 #include <interrupt.h>
 #include <mm/frame.h>
 #include <mm/page.h>
 #include <genarch/mm/page_pt.h>
 #include <mm/tlb.h>
 #include <mm/as.h>
 #include <mm/slab.h>
 #include <synch/waitq.h>
 #include <synch/futex.h>
 #include <arch/arch.h>
 #include <arch.h>
 #include <arch/faddr.h>
 #include <ipc/ipc.h>
 #include <macros.h>
 #include <adt/btree.h>
-#include <console/klog.h>
 #include <smp/smp.h>
 #include <ddi/ddi.h>
+#include <main/main.h>
 /** Global configuration structure. */
 config_t config;
 /** Initial user-space tasks */
 init_t init = {
     .cnt = 0
 };
 /** Boot allocations. */
 ballocs_t ballocs = {
     .base = NULL,
     .size = 0
 };
 context_t ctx;
 /*
  * These 'hardcoded' variables will be intialized by
  * the linker or the low level assembler code with
  * appropriate sizes and addresses.
  */
-/**< Virtual address of where the kernel is loaded. */
+/** Virtual address of where the kernel is loaded. */
 uintptr_t hardcoded_load_address = 0;
-/**< Size of the kernel code in bytes. */
+/** Size of the kernel code in bytes. */
 size_t hardcoded_ktext_size = 0;
-/**< Size of the kernel data in bytes. */
+/** Size of the kernel data in bytes. */
 size_t hardcoded_kdata_size = 0;
-/**< Lowest safe stack virtual address. */
+/** Lowest safe stack virtual address. */
 uintptr_t stack_safe = 0;
-void main_bsp(void);
-void main_ap(void);
 /*
  * These two functions prevent stack from underflowing during the
  * kernel boot phase when SP is set to the very top of the reserved
  * space. The stack could get corrupted by a fooled compiler-generated
  * pop sequence otherwise.
  */
 static void main_bsp_separated_stack(void);
 #ifdef CONFIG_SMP
 static void main_ap_separated_stack(void);
 #endif
 #define CONFIG_STACK_SIZE   ((1 << STACK_FRAMES) * STACK_SIZE)
 /** Main kernel routine for bootstrap CPU.
+ *
- * Initializes the kernel by bootstrap CPU.
+ * The code here still runs on the boot stack, which knows nothing about
- * This function passes control directly to
+ * preemption counts.  Because of that, this function cannot directly call
+ * functions that disable or enable preemption (e.g. spinlock_lock()). The
+ * primary task of this function is to calculate address of a new stack and
- * main_bsp_separated_stack().
+ * switch to it.
+ *
  * Assuming interrupts_disable().
+ *
  */
 void main_bsp(void)
+{
     config.cpu_count = 1;
     config.cpu_active = 1;
     config.base = hardcoded_load_address;
     config.kernel_size = ALIGN_UP(hardcoded_ktext_size +
         hardcoded_kdata_size, PAGE_SIZE);
     config.stack_size = CONFIG_STACK_SIZE;
     /* Initialy the stack is placed just after the kernel */
     config.stack_base = config.base + config.kernel_size;
     /* Avoid placing stack on top of init */
     count_t i;
     for (i = 0; i < init.cnt; i++) {
         if (PA_overlaps(config.stack_base, config.stack_size,
             init.tasks[i].addr, init.tasks[i].size))
             config.stack_base = ALIGN_UP(init.tasks[i].addr +
                 init.tasks[i].size, config.stack_size);
+    }
     /* Avoid placing stack on top of boot allocations. */
     if (ballocs.size) {
         if (PA_overlaps(config.stack_base, config.stack_size,
             ballocs.base, ballocs.size))
             config.stack_base = ALIGN_UP(ballocs.base +
                 ballocs.size, PAGE_SIZE);
+    }
     if (config.stack_base < stack_safe)
         config.stack_base = ALIGN_UP(stack_safe, PAGE_SIZE);
     context_save(&ctx);
     context_set(&ctx, FADDR(main_bsp_separated_stack), config.stack_base,
         THREAD_STACK_SIZE);
     context_restore(&ctx);
     /* not reached */
+}
 /** Main kernel routine for bootstrap CPU using new stack.
+ *
  * Second part of main_bsp().
+ *
  */
 void main_bsp_separated_stack(void)
+{
-    task_t *k;
-    thread_t *t;
+    /* Keep this the first thing. */
-    count_t i;
     the_initialize(THE);
+    version_print();
+    LOG("\nconfig.base=%#" PRIp " config.kernel_size=%" PRIs
+        "\nconfig.stack_base=%#" PRIp " config.stack_size=%" PRIs,
+        config.base, config.kernel_size, config.stack_base,
+        config.stack_size);
+#ifdef CONFIG_KCONSOLE
     /*
      * kconsole data structures must be initialized very early
      * because other subsystems will register their respective
      * commands.
      */
-    kconsole_init();
+    LOG_EXEC(kconsole_init());
+#endif
     /*
      * Exception handler initialization, before architecture
      * starts adding its own handlers
      */
-    exc_init();
+    LOG_EXEC(exc_init());
     /*
      * Memory management subsystems initialization.
      */
-    arch_pre_mm_init();
+    LOG_EXEC(arch_pre_mm_init());
-    frame_init();
+    LOG_EXEC(frame_init());
-    /* Initialize at least 1 memory segment big enough for slab to work. */
-    slab_cache_init();
-    btree_init();
-    as_init();
-    page_init();
-    tlb_init();
-    ddi_init();
-    tasklet_init();
-    arch_post_mm_init();
+    /* Initialize at least 1 memory segment big enough for slab to work. */
+    LOG_EXEC(slab_cache_init());
+    LOG_EXEC(btree_init());
-    version_print();
+    LOG_EXEC(as_init());
-    printf("kernel: %.*p hardcoded_ktext_size=%zd KB, "
+    LOG_EXEC(page_init());
+    LOG_EXEC(tlb_init());
-        "hardcoded_kdata_size=%zd KB\n", sizeof(uintptr_t) * 2,
+    LOG_EXEC(ddi_init());
-        config.base, SIZE2KB(hardcoded_ktext_size),
+    LOG_EXEC(tasklet_init());
-        SIZE2KB(hardcoded_kdata_size));
+    LOG_EXEC(arch_post_mm_init());
-    printf("stack:  %.*p size=%zd KB\n", sizeof(uintptr_t) * 2,
+    LOG_EXEC(arch_pre_smp_init());
-        config.stack_base, SIZE2KB(config.stack_size));
+    LOG_EXEC(smp_init());
-    arch_pre_smp_init();
-    smp_init();
     /* Slab must be initialized after we know the number of processors. */
-    slab_enable_cpucache();
+    LOG_EXEC(slab_enable_cpucache());
-    printf("Detected %zu CPU(s), %llu MB free memory\n",
+    printf("Detected %" PRIc " CPU(s), %" PRIu64" MiB free memory\n",
         config.cpu_count, SIZE2MB(zone_total_size()));
-    cpu_init();
+    LOG_EXEC(cpu_init());
-    calibrate_delay_loop();
+    LOG_EXEC(calibrate_delay_loop());
-    clock_counter_init();
+    LOG_EXEC(clock_counter_init());
-    timeout_init();
+    LOG_EXEC(timeout_init());
-    scheduler_init();
+    LOG_EXEC(scheduler_init());
-    task_init();
+    LOG_EXEC(task_init());
-    thread_init();
+    LOG_EXEC(thread_init());
-    futex_init();
+    LOG_EXEC(futex_init());
-    klog_init();
     if (init.cnt > 0) {
+        count_t i;
         for (i = 0; i < init.cnt; i++)
-            printf("init[%zd].addr=%.*p, init[%zd].size=%zd\n", i,
+            LOG("init[%" PRIc "].addr=%#" PRIp ", init[%" PRIc
-                sizeof(uintptr_t) * 2, init.tasks[i].addr, i,
+                "].size=%#" PRIs "\n", i, init.tasks[i].addr, i,
                 init.tasks[i].size);
     } else
         printf("No init binaries found\n");
-    ipc_init();
+    LOG_EXEC(ipc_init());
+    LOG_EXEC(klog_init());
+    LOG_EXEC(console_init());
+#ifdef CONFIG_KCONSOLE
+    LOG_EXEC(kconsole_notify_init());
+#endif
     /*
      * Create kernel task.
      */
-    k = task_create(AS_KERNEL, "kernel");
+    task_t *kernel = task_create(AS_KERNEL, "kernel");
-    if (!k)
+    if (!kernel)
-        panic("can't create kernel task\n");
+        panic("Cannot create kernel task.");
     /*
      * Create the first thread.
      */
+    thread_t *kinit_thread
-    t = thread_create(kinit, NULL, k, 0, "kinit", true);
+        = thread_create(kinit, NULL, kernel, 0, "kinit", true);
-    if (!t)
+    if (!kinit_thread)
-        panic("can't create kinit thread\n");
+        panic("Cannot create kinit thread.");
-    thread_ready(t);
+    LOG_EXEC(thread_ready(kinit_thread));
     /*
      * This call to scheduler() will return to kinit,
      * starting the thread of kernel threads.
      */
     scheduler();
     /* not reached */
+}
 #ifdef CONFIG_SMP
 /** Main kernel routine for application CPUs.
+ *
  * Executed by application processors, temporary stack
  * is at ctx.sp which was set during BSP boot.
  * This function passes control directly to
  * main_ap_separated_stack().
+ *
  * Assuming interrupts_disable()'d.
+ *
  */
 void main_ap(void)
+{
     /*
      * Incrementing the active CPU counter will guarantee that the
      * *_init() functions can find out that they need to
      * do initialization for AP only.
      */
     config.cpu_active++;
     /*
      * The THE structure is well defined because ctx.sp is used as stack.
      */
     the_initialize(THE);
     arch_pre_mm_init();
     frame_init();
     page_init();
     tlb_init();
     arch_post_mm_init();
     cpu_init();
     calibrate_delay_loop();
     arch_post_cpu_init();
     the_copy(THE, (the_t *) CPU->stack);
     /*
      * If we woke kmp up before we left the kernel stack, we could
      * collide with another CPU coming up. To prevent this, we
      * switch to this cpu's private stack prior to waking kmp up.
      */
+    context_save(&CPU->saved_context);
     context_set(&CPU->saved_context, FADDR(main_ap_separated_stack),
         (uintptr_t) CPU->stack, CPU_STACK_SIZE);
     context_restore(&CPU->saved_context);
     /* not reached */
+}
 /** Main kernel routine for application CPUs using new stack.
+ *
  * Second part of main_ap().
+ *
  */
 void main_ap_separated_stack(void)
+{
     /*
      * Configure timeouts for this cpu.
      */
     timeout_init();
     waitq_wakeup(&ap_completion_wq, WAKEUP_FIRST);
     scheduler();
     /* not reached */
+}
 #endif /* CONFIG_SMP */
 /** @}
  */

Subversion Repositories HelenOS

(root)/branches/dd/kernel/generic/src/main/main.c – Rev 3022 → 4055