Palacios Public Git Repository

To checkout Palacios execute

  git clone http://v3vee.org/palacios/palacios.web/palacios.git
This will give you the master branch. You probably want the devel branch or one of the release branches. To switch to the devel branch, simply execute
  cd palacios
  git checkout --track -b devel origin/devel
The other branches are similar.


added configurable host OS thread support
[palacios.git] / palacios / src / palacios / vmm_config.c
index 62f4de5..6ffa3e4 100644 (file)
+ /* 
+ * This file is part of the Palacios Virtual Machine Monitor developed
+ * by the V3VEE Project with funding from the United States National 
+ * Science Foundation and the Department of Energy.  
+ *
+ * The V3VEE Project is a joint project between Northwestern University
+ * and the University of New Mexico.  You can find out more at 
+ * http://www.v3vee.org
+ *
+ * Copyright (c) 2008, Jack Lange <jarusl@cs.northwestern.edu> 
+ * Copyright (c) 2008, The V3VEE Project <http://www.v3vee.org> 
+ * All rights reserved.
+ *
+ * Author: Jack Lange <jarusl@cs.northwestern.edu>
+ *
+ * This is free software.  You are permitted to use,
+ * redistribute, and modify it as specified in the file "V3VEE_LICENSE".
+ */
+
 #include <palacios/vmm_config.h>
 #include <palacios/vmm.h>
 #include <palacios/vmm_debug.h>
+#include <palacios/vmm_msr.h>
+#include <palacios/vmm_decoder.h>
+#include <palacios/vmm_telemetry.h>
+#include <palacios/vmm_mem.h>
+#include <palacios/vmm_hypercall.h>
+#include <palacios/vmm_dev_mgr.h>
+#include <palacios/vmm_cpuid.h>
+#include <palacios/vmm_xml.h>
+#include <palacios/vmm_io.h>
+#include <palacios/vmm_msr.h>
+#include <palacios/vmm_mptable.h>
 
-#include <devices/serial.h>
-#include <devices/keyboard.h>
-#include <devices/8259a.h>
-#include <devices/8254.h>
-#include <devices/nvram.h>
-#include <devices/generic.h>
 
 
 
-int config_guest(struct guest_info * info, void * config_ptr) {
 
-  struct guest_mem_layout * layout = (struct guest_mem_layout *)config_ptr;
-  extern v3_cpu_arch_t v3_cpu_type;
-  void * region_start;
 
-  PrintDebug("Time Init\n");
-  v3_init_time(info);
+#include <palacios/vmm_host_events.h>
+#include <palacios/vmm_socket.h>
 
-  PrintDebug("Shadow map Init\n");
-  init_shadow_map(info);
-  
-  if (v3_cpu_type == V3_SVM_REV3_CPU) {
-    info->shdw_pg_mode = NESTED_PAGING;
-  } else {
-    init_shadow_page_state(info);
-    info->shdw_pg_mode = SHADOW_PAGING;
-  }
-  
-  info->cpu_mode = REAL;
-  info->mem_mode = PHYSICAL_MEM;
-  
-  init_vmm_io_map(info);
-  init_interrupt_state(info);
-  
-  dev_mgr_init(info);
-  
-  
-  int i;
+#include "vmm_config_class.h"
 
-  
-  //     SerialPrint("Guest Mem Dump at 0x%x\n", 0x100000);
-  //PrintDebugMemDump((unsigned char *)(0x100000), 261 * 1024);
-  if (layout->magic != MAGIC_CODE) {
-    
-    PrintDebug("Layout Magic Mismatch (0x%x)\n", layout->magic);
-    return -1;
-  }
-  
-  PrintDebug("%d layout regions\n", layout->num_regions);
-  
-  region_start = (void *)&(layout->regions[layout->num_regions]);
-  
-  PrintDebug("region start = 0x%x\n", region_start);
-  
-  for (i = 0; i < layout->num_regions; i++) {
-    struct layout_region * reg = &(layout->regions[i]);
-    uint_t num_pages = (reg->length / PAGE_SIZE) + ((reg->length % PAGE_SIZE) ? 1 : 0);
-    void * guest_mem = V3_AllocPages(num_pages);
-    
-    PrintDebug("Layout Region %d bytes\n", reg->length);
-    memcpy(guest_mem, region_start, reg->length);
-    
-    PrintDebugMemDump((unsigned char *)(guest_mem), 16);
-    
-    add_shadow_region_passthrough(info, reg->final_addr, reg->final_addr + (num_pages * PAGE_SIZE), (addr_t)guest_mem);
-    
-    PrintDebug("Adding Shadow Region (0x%x-0x%x) -> 0x%x\n", reg->final_addr, reg->final_addr + (num_pages * PAGE_SIZE), guest_mem);
+// This is used to access the configuration file index table
+struct file_hdr {
+    uint32_t index;
+    uint32_t size;
+    uint64_t offset;
+};
+
+struct file_idx_table {
+    uint64_t num_files;
+    struct file_hdr hdrs[0];
+};
+
+
+
+
+static int setup_memory_map(struct v3_vm_info * vm, v3_cfg_tree_t * cfg);
+static int setup_devices(struct v3_vm_info * vm, v3_cfg_tree_t * cfg);
+
+
+
+char * v3_cfg_val(v3_cfg_tree_t * tree, char * tag) {
+    char * attrib = (char *)v3_xml_attr(tree, tag);
+    v3_cfg_tree_t * child_entry = v3_xml_child(tree, tag);
+    char * val = NULL;
+
+    if ((child_entry != NULL) && (attrib != NULL)) {
+       PrintError("Duplicate Configuration parameters present for %s\n", tag);
+       return NULL;
+    }
+
+    if (attrib == NULL) {
+       val = v3_xml_txt(child_entry);
+       
+       if ( val[0] == 0 )
+               val = NULL;
+    } else {
+       val = attrib;
+    }
     
-    region_start += reg->length;
-  }
-  
-      //     
-  add_shadow_region_passthrough(info, 0x0, 0xa0000, (addr_t)V3_AllocPages(160));
-  
-  add_shadow_region_passthrough(info, 0xa0000, 0xc0000, 0xa0000); 
-  //hook_guest_mem(info, 0xa0000, 0xc0000, passthrough_mem_read, passthrough_mem_write, NULL);
-  
-  
-  // TEMP
-  //add_shadow_region_passthrough(info, 0xc0000, 0xc8000, 0xc0000);
-  
-  if (1) {
-    add_shadow_region_passthrough(info, 0xc7000, 0xc8000, (addr_t)V3_AllocPages(1));
-    if (add_shadow_region_passthrough(info, 0xc8000, 0xf0000, (addr_t)V3_AllocPages(40)) == -1) {
-      PrintDebug("Error adding shadow region\n");
+    return val;
+}
+
+v3_cfg_tree_t * v3_cfg_subtree(v3_cfg_tree_t * tree, char * tag) {
+    return v3_xml_child(tree, tag);
+}
+
+v3_cfg_tree_t * v3_cfg_next_branch(v3_cfg_tree_t * tree) {
+    return v3_xml_next(tree);
+}
+
+
+
+struct v3_cfg_file * v3_cfg_get_file(struct v3_vm_info * vm, char * tag) {
+    struct v3_cfg_file * file = NULL;
+
+    file = (struct v3_cfg_file *)v3_htable_search(vm->cfg_data->file_table, (addr_t)tag);
+
+    return file;
+}
+
+
+static uint_t file_hash_fn(addr_t key) {
+    char * name = (char *)key;
+    return v3_hash_buffer((uchar_t *)name, strlen(name));
+}
+
+static int file_eq_fn(addr_t key1, addr_t key2) {
+    char * name1 = (char *)key1;
+    char * name2 = (char *)key2;
+
+    return (strcmp(name1, name2) == 0);
+}
+
+static struct v3_config * parse_config(void * cfg_blob) {
+    struct v3_config * cfg = NULL;
+    int offset = 0;
+    uint_t xml_len = 0; 
+    struct file_idx_table * files = NULL;
+    v3_cfg_tree_t * file_tree = NULL;
+
+    V3_Print("cfg data at %p\n", cfg_blob);
+
+    if (memcmp(cfg_blob, "v3vee\0\0\0", 8) != 0) {
+       PrintError("Invalid Configuration Header\n");
+       return NULL;
     }
-  } else {
-    add_shadow_region_passthrough(info, 0xc0000, 0xc8000, 0xc0000);
-    add_shadow_region_passthrough(info, 0xc8000, 0xf0000, 0xc8000);
-  }
-  
-  
-  //add_shadow_region_passthrough(info, 0x100000, 0x2000000, (addr_t)Allocate_VMM_Pages(8192));
-  add_shadow_region_passthrough(info, 0x100000, 0x1000000, (addr_t)V3_AllocPages(4096));
-  
-  add_shadow_region_passthrough(info, 0x1000000, 0x8000000, (addr_t)V3_AllocPages(32768));
-  
-  // test - give linux accesss to PCI space - PAD
-  add_shadow_region_passthrough(info, 0xc0000000,0xffffffff,0xc0000000);
-  
-  
-  print_shadow_map(&(info->mem_map));
 
-  
-  {
+    offset += 8;
+
+    cfg = (struct v3_config *)V3_Malloc(sizeof(struct v3_config));
+    memset(cfg, 0, sizeof(struct v3_config));
+
+    cfg->blob = cfg_blob;
+    INIT_LIST_HEAD(&(cfg->file_list));
+    cfg->file_table = v3_create_htable(0, file_hash_fn, file_eq_fn);
     
-    struct vm_device * nvram = create_nvram();
-    //struct vm_device * timer = create_timer();
-    struct vm_device * pic = create_pic();
-    struct vm_device * keyboard = create_keyboard();
-    struct vm_device * pit = create_pit(); 
-    //struct vm_device * serial = create_serial();
+    xml_len = *(uint32_t *)(cfg_blob + offset);
+    offset += 4;
+
+    cfg->cfg = (v3_cfg_tree_t *)v3_xml_parse((uint8_t *)(cfg_blob + offset));
+    offset += xml_len;
+   
+    offset += 8;
+
+    files = (struct file_idx_table *)(cfg_blob + offset);
+
+    V3_Print("Number of files in cfg: %d\n", (uint32_t)(files->num_files));
+
+    file_tree = v3_cfg_subtree(v3_cfg_subtree(cfg->cfg, "files"), "file");
+
+    while (file_tree) {
+       char * id = v3_cfg_val(file_tree, "id");
+       char * index = v3_cfg_val(file_tree, "index");
+       int idx = atoi(index);
+       struct file_hdr * hdr = &(files->hdrs[idx]);
+       struct v3_cfg_file * file = NULL;
+
+       file = (struct v3_cfg_file *)V3_Malloc(sizeof(struct v3_cfg_file));
+       
+       if (!file) {
+           PrintError("Could not allocate file structure\n");
+           return NULL;
+       }
+
+
+       V3_Print("File index=%d id=%s\n", idx, id);
+
+       strncpy(file->tag, id, 256);
+       file->size = hdr->size;
+       file->data = cfg_blob + hdr->offset;
+
+       V3_Print("Storing file data offset = %d, size=%d\n", (uint32_t)hdr->offset, hdr->size);
+       V3_Print("file data at %p\n", file->data);
+       list_add( &(file->file_node), &(cfg->file_list));
+
+       V3_Print("Keying file to name\n");
+       v3_htable_insert(cfg->file_table, (addr_t)(file->tag), (addr_t)(file));
+
+       V3_Print("Iterating to next file\n");
+
+       file_tree = v3_cfg_next_branch(file_tree);
+    }
+
+    V3_Print("Configuration parsed successfully\n");
+
+    return cfg;
+}
+
+
+static inline uint32_t get_alignment(char * align_str) {
+    // default is 4KB alignment
+    uint32_t alignment = PAGE_SIZE_4KB;
+
+    if (align_str != NULL) {
+       if (strcasecmp(align_str, "2MB") == 0) {
+           alignment = PAGE_SIZE_2MB;
+       } else if (strcasecmp(align_str, "4MB") == 0) {
+           alignment = PAGE_SIZE_4MB;
+       }
+    }
     
+#ifndef CONFIG_ALIGNED_PG_ALLOC
+    if (alignment != PAGE_SIZE_4KB) {
+       PrintError("Aligned page allocations are not supported in this host (requested alignment=%d)\n", alignment);
+       PrintError("Ignoring alignment request\n");
+    }
+#endif 
+
+    return alignment;
+}
+static int pre_config_vm(struct v3_vm_info * vm, v3_cfg_tree_t * vm_cfg) {
+
+
+    char * memory_str = v3_cfg_val(vm_cfg, "memory");
+    char * schedule_hz_str = v3_cfg_val(vm_cfg, "schedule_hz");
+    char * vm_class = v3_cfg_val(vm_cfg, "class");
+    char * align_str = v3_cfg_val(v3_cfg_subtree(vm_cfg, "memory"), "alignment");
+    uint32_t sched_hz = 100;   // set the schedule frequency to 100 HZ
     
-#define GENERIC 1
+    if (!memory_str) {
+       PrintError("Memory is a required configuration parameter\n");
+       return -1;
+    }
     
-#if GENERIC
-    generic_port_range_type range[] = {
-#if 1
-      {0x00, 0x07, GENERIC_PRINT_AND_IGNORE},   // DMA 1 channels 0,1,2,3 (address, counter)
-      {0xc0, 0xc7, GENERIC_PRINT_AND_IGNORE},   // DMA 2 channels 4,5,6,7 (address, counter)
-      {0x87, 0x87, GENERIC_PRINT_AND_IGNORE},   // DMA 1 channel 0 page register
-      {0x83, 0x83, GENERIC_PRINT_AND_IGNORE},   // DMA 1 channel 1 page register
-      {0x81, 0x81, GENERIC_PRINT_AND_IGNORE},   // DMA 1 channel 2 page register
-      {0x82, 0x82, GENERIC_PRINT_AND_IGNORE},   // DMA 1 channel 3 page register
-      {0x8f, 0x8f, GENERIC_PRINT_AND_IGNORE},   // DMA 2 channel 4 page register
-      {0x8b, 0x8b, GENERIC_PRINT_AND_IGNORE},   // DMA 2 channel 5 page register
-      {0x89, 0x89, GENERIC_PRINT_AND_IGNORE},   // DMA 2 channel 6 page register
-      {0x8a, 0x8a, GENERIC_PRINT_AND_IGNORE},   // DMA 2 channel 7 page register
-      {0x08, 0x0f, GENERIC_PRINT_AND_IGNORE},   // DMA 1 misc registers (csr, req, smask,mode,clearff,reset,enable,mmask)
-      {0xd0, 0xde, GENERIC_PRINT_AND_IGNORE},   // DMA 2 misc registers
-#endif
-      
-      
-      {0x3f8, 0x3f8+7, GENERIC_PRINT_AND_IGNORE},      // COM 1
-      {0x2f8, 0x2f8+7, GENERIC_PRINT_AND_IGNORE},      // COM 2
-      {0x3e8, 0x3e8+7, GENERIC_PRINT_AND_IGNORE},      // COM 3
-      {0x2e8, 0x2e8+7, GENERIC_PRINT_AND_IGNORE},      // COM 4
-      
-#if 0
-      {0x170, 0x178, GENERIC_PRINT_AND_PASSTHROUGH}, // IDE 1
-      {0x376, 0x377, GENERIC_PRINT_AND_PASSTHROUGH}, // IDE 1
-      {0x1f0, 0x1f8, GENERIC_PRINT_AND_PASSTHROUGH}, // IDE 0
-      {0x3f6, 0x3f7, GENERIC_PRINT_AND_PASSTHROUGH}, // IDE 0
-#endif
-      
-
-#if 0
-      {0x3f0, 0x3f2, GENERIC_PRINT_AND_IGNORE}, // Primary floppy controller (base,statusa/statusb,DOR)
-      {0x3f4, 0x3f5, GENERIC_PRINT_AND_IGNORE}, // Primary floppy controller (mainstat/datarate,data)
-      {0x3f7, 0x3f7, GENERIC_PRINT_AND_IGNORE}, // Primary floppy controller (DIR)
-      {0x370, 0x372, GENERIC_PRINT_AND_IGNORE}, // Secondary floppy controller (base,statusa/statusb,DOR)
-      {0x374, 0x375, GENERIC_PRINT_AND_IGNORE}, // Secondary floppy controller (mainstat/datarate,data)
-      {0x377, 0x377, GENERIC_PRINT_AND_IGNORE}, // Secondary floppy controller (DIR)
-      
+    PrintDebug("Memory=%s\n", memory_str);
+    if (align_str) {
+        PrintDebug("Alignment=%s\n", align_str);
+    } else {
+        PrintDebug("Alignment defaulted to 4KB.\n");
+    }
+
+    // Amount of ram the Guest will have, always in MB
+    vm->mem_size = (addr_t)atoi(memory_str) * 1024 * 1024;
+    vm->mem_align = get_alignment(align_str);
+
+    PrintDebug("Alignment for %lu bytes of memory computed as 0x%x\n", vm->mem_size, vm->mem_align);
+
+    if (strcasecmp(vm_class, "PC") == 0) {
+       vm->vm_class = V3_PC_VM;
+    } else {
+       PrintError("Invalid VM class\n");
+       return -1;
+    }
+
+#ifdef CONFIG_TELEMETRY
+    {
+       char * telemetry = v3_cfg_val(vm_cfg, "telemetry");
+
+       // This should go first, because other subsystems will depend on the guest_info flag    
+       if ((telemetry) && (strcasecmp(telemetry, "enable") == 0)) {
+           vm->enable_telemetry = 1;
+       } else {
+           vm->enable_telemetry = 0;
+       }
+    }
 #endif
 
-      //         {0x378, 0x400, GENERIC_PRINT_AND_IGNORE}
-      
-      {0,0,0},  // sentinal - must be last
-      
-    };
-    
+    if (v3_init_vm(vm) == -1) {
+       PrintError("Failed to initialize VM\n");
+       return -1;
+    }
 
-    struct vm_device * generic = create_generic(range, NULL, NULL);
+
+
+   if (schedule_hz_str) {
+       sched_hz = atoi(schedule_hz_str);
+    }
+
+    PrintDebug("CPU_KHZ = %d, schedule_freq=%p\n", V3_CPU_KHZ(), 
+              (void *)(addr_t)sched_hz);
+
+    vm->yield_cycle_period = (V3_CPU_KHZ() * 1000) / sched_hz;
     
-#endif
+    return 0;
+}
+
+static int determine_paging_mode(struct guest_info *info, v3_cfg_tree_t * core_cfg)
+{
+    extern v3_cpu_arch_t v3_cpu_types[];
+
+    v3_cfg_tree_t * vm_tree = info->vm_info->cfg_data->cfg;
+    v3_cfg_tree_t * pg_tree = v3_cfg_subtree(vm_tree, "paging");
+    char * pg_mode          = v3_cfg_val(pg_tree, "mode");
+    char * page_size        = v3_cfg_val(pg_tree, "page_size");
     
-    v3_attach_device(info, nvram);
-    //v3_attach_device(info, timer);
-    v3_attach_device(info, pic);
-    v3_attach_device(info, pit);
-    v3_attach_device(info, keyboard);
-    // v3_attach_device(info, serial);
+    PrintDebug("Paging mode specified as %s\n", pg_mode);
+
+    if (pg_mode) {
+       if ((strcasecmp(pg_mode, "nested") == 0)) {
+           if (v3_cpu_types[info->cpu_id] == V3_SVM_REV3_CPU) {
+               info->shdw_pg_mode = NESTED_PAGING;
+           } else {
+               PrintError("Nested paging not supported on this hardware. Defaulting to shadow paging\n");
+               info->shdw_pg_mode = SHADOW_PAGING;
+           }
+       } else if ((strcasecmp(pg_mode, "shadow") == 0)) {
+           info->shdw_pg_mode = SHADOW_PAGING;
+       } else {
+           PrintError("Invalid paging mode (%s) specified in configuration. Defaulting to shadow paging\n", pg_mode);
+           info->shdw_pg_mode = SHADOW_PAGING;
+       }
+    } else {
+       PrintDebug("No paging type specified in configuration. Defaulting to shadow paging\n");
+       info->shdw_pg_mode = SHADOW_PAGING;
+    }
+
+
+    if (info->shdw_pg_mode == NESTED_PAGING) {
+       PrintDebug("Guest Paging Mode: NESTED_PAGING\n");
+       if (strcasecmp(page_size, "4kb") == 0) { /* TODO: this may not be an ideal place for this */
+           info->vm_info->paging_size = PAGING_4KB;
+       } else if (strcasecmp(page_size, "2mb") == 0) {
+           info->vm_info->paging_size = PAGING_2MB;
+       } else {
+           PrintError("Invalid VM paging size: '%s'\n", page_size);
+           return -1;
+       }
+       PrintDebug("VM page size=%s\n", page_size);
+    } else if (info->shdw_pg_mode == SHADOW_PAGING) {
+        PrintDebug("Guest Paging Mode: SHADOW_PAGING\n");
+    } else {
+       PrintError("Guest paging mode incorrectly set.\n");
+       return -1;
+    }
+
+    if (v3_cfg_val(pg_tree, "large_pages") != NULL) {
+       if (strcasecmp(v3_cfg_val(pg_tree, "large_pages"), "true") == 0) {
+           info->use_large_pages = 1;
+           PrintDebug("Use of large pages in memory virtualization enabled.\n");
+       }
+    }
+    return 0;
+}
+
+static int pre_config_core(struct guest_info * info, v3_cfg_tree_t * core_cfg) {
+
+    if (determine_paging_mode(info, core_cfg))
+       return -1;
+
+    v3_init_core(info);
+
+    if (info->vm_info->vm_class == V3_PC_VM) {
+       if (pre_config_pc_core(info, core_cfg) == -1) {
+           PrintError("PC Post configuration failure\n");
+           return -1;
+       }
+    } else {
+       PrintError("Invalid VM Class\n");
+       return -1;
+    }
+
+    return 0;
+}
+
 
 
-#if GENERIC
-    // Important that this be attached last!
-    v3_attach_device(info, generic);
+static int post_config_vm(struct v3_vm_info * vm, v3_cfg_tree_t * cfg) {
     
-#endif
+    vm->run_state = VM_STOPPED;
+
+    // Configure the memory map for the guest
+    if (setup_memory_map(vm, cfg) == -1) {
+       PrintError("Setting up guest memory map failed...\n");
+       return -1;
+    }
     
-    PrintDebugDevMgr(info);
-  }
-  
-  // give keyboard interrupts to vm
-  // no longer needed since we have a keyboard device
-  //hook_irq(&vm_info, 1);
-  
-#if 1
-  // give floppy controller to vm
-  v3_hook_irq_for_guest_injection(info, 6);
-#endif
-  
-#if 1
-  //primary ide
-  v3_hook_irq_for_guest_injection(info, 14);
-  
-  // secondary ide
-  v3_hook_irq_for_guest_injection(info, 15);
-#endif
+    //v3_hook_io_port(info, 1234, &IO_Read, NULL, info);
   
+    if (setup_devices(vm, cfg) == -1) {
+       PrintError("Failed to setup devices\n");
+       return -1;
+    }
 
-  info->rip = 0xfff0;
-  info->vm_regs.rsp = 0x0;
-  
 
-  return 0;
+    //    v3_print_io_map(info);
+    v3_print_msr_map(vm);
+
+
+    if (vm->vm_class == V3_PC_VM) {
+       if (post_config_pc(vm, cfg) == -1) {
+           PrintError("PC Post configuration failure\n");
+           return -1;
+       }
+    } else {
+       PrintError("Invalid VM Class\n");
+       return -1;
+    }
+
+    return 0;
 }
 
 
 
+static int post_config_core(struct guest_info * info, v3_cfg_tree_t * cfg) {
 
+    info->core_run_state = CORE_STOPPED;
+    if (info->vm_info->vm_class == V3_PC_VM) {
+       if (post_config_pc_core(info, cfg) == -1) {
+           PrintError("PC Post configuration failure\n");
+           return -1;
+       }
+    } else {
+       PrintError("Invalid VM Class\n");
+       return -1;
+    }
 
 
+    return 0;
+}
 
 
 
+static struct v3_vm_info * allocate_guest(int num_cores) {
+    int guest_state_size = sizeof(struct v3_vm_info) + (sizeof(struct guest_info) * num_cores);
+    struct v3_vm_info * vm = V3_Malloc(guest_state_size);
 
+    memset(vm, 0, guest_state_size);
 
+    vm->num_cores = num_cores;
 
+    return vm;
+}
 
 
 
+struct v3_vm_info * v3_config_guest(void * cfg_blob) {
+    v3_cpu_arch_t cpu_type = v3_get_cpu_type(V3_Get_CPU());
+    struct v3_config * cfg_data = NULL;
+    struct v3_vm_info * vm = NULL;
+    int num_cores = 0;
+    int i = 0;
+    v3_cfg_tree_t * cores_cfg = NULL;
+    v3_cfg_tree_t * per_core_cfg = NULL;
 
+    if (cpu_type == V3_INVALID_CPU) {
+       PrintError("Configuring guest on invalid CPU\n");
+       return NULL;
+    }
+
+    cfg_data = parse_config(cfg_blob);
+
+    if (!cfg_data) {
+       PrintError("Could not parse configuration\n");
+       return NULL;
+    }
 
+    cores_cfg = v3_cfg_subtree(cfg_data->cfg, "cores");
 
+    if (!cores_cfg) {
+       PrintError("Could not find core configuration (new config format required)\n");
+       return NULL;
+    }
 
+    num_cores = atoi(v3_cfg_val(cores_cfg, "count"));
 
+    if (num_cores == 0) {
+       PrintError("No cores specified in configuration\n");
+       return NULL;
+    }
 
+    V3_Print("Configuring %d cores\n", num_cores);
 
-#if 0
+    vm = allocate_guest(num_cores);    
+
+    if (!vm) {
+       PrintError("Could not allocate %d core guest\n", vm->num_cores);
+       return NULL;
+    }
+
+    vm->cfg_data = cfg_data;
+
+    V3_Print("Preconfiguration\n");
+
+    if (pre_config_vm(vm, vm->cfg_data->cfg) == -1) {
+       PrintError("Error in preconfiguration\n");
+       return NULL;
+    }
+
+
+    V3_Print("Per core configuration\n");
+    per_core_cfg = v3_cfg_subtree(cores_cfg, "core");
+
+    // per core configuration
+    for (i = 0; i < vm->num_cores; i++) {
+       struct guest_info * info = &(vm->cores[i]);
+
+       info->cpu_id = i;
+       info->vm_info = vm;
+       info->core_cfg_data = per_core_cfg;
+
+       if (pre_config_core(info, per_core_cfg) == -1) {
+           PrintError("Error in core %d preconfiguration\n", i);
+           return NULL;
+       }
+
+       per_core_cfg = v3_cfg_next_branch(per_core_cfg);
+    }
+
+
+    V3_Print("Post Configuration\n");
+
+    if (post_config_vm(vm, vm->cfg_data->cfg) == -1) {
+       PrintError("Error in postconfiguration\n");
+       return NULL;
+    }
+
+
+    per_core_cfg = v3_cfg_subtree(cores_cfg, "core");
+
+    // per core configuration
+    for (i = 0; i < vm->num_cores; i++) {
+       struct guest_info * info = &(vm->cores[i]);
+
+       post_config_core(info, per_core_cfg);
+
+       per_core_cfg = v3_cfg_next_branch(per_core_cfg);
+    }
+
+    V3_Print("Configuration successfull\n");
+
+    return vm;
+}
+
+
+
+
+
+static int setup_memory_map(struct v3_vm_info * vm, v3_cfg_tree_t * cfg) {
+    v3_cfg_tree_t * mem_region = v3_cfg_subtree(v3_cfg_subtree(cfg, "memmap"), "region");
+
+    while (mem_region) {
+       addr_t start_addr = atox(v3_cfg_val(mem_region, "start"));
+       addr_t end_addr = atox(v3_cfg_val(mem_region, "end"));
+       addr_t host_addr = atox(v3_cfg_val(mem_region, "host_addr"));
 
     
-    if (0) {
-      
-      //    add_shared_mem_range(&(vm_info.mem_layout), 0, 0x800000, 0x10000);    
-      //    add_shared_mem_range(&(vm_info.mem_layout), 0, 0x1000000, 0);
-      
-      rip = (ulong_t)(void*)&BuzzVM;
-      //  rip -= 0x10000;
-      //    rip = (addr_t)(void*)&exit_test;
-      //  rip -= 0x2000;
-      vm_info.rip = rip;
-      rsp = (addr_t)Alloc_Page();
-      
-      vm_info.vm_regs.rsp = (rsp +4092 );// - 0x2000;
-      
-            
-    } else if (0) {
-      //add_shared_mem_range(&(vm_info.mem_layout), 0x0, 0x1000, 0x100000);
-      //      add_shared_mem_range(&(vm_info.mem_layout), 0x0, 0x100000, 0x0);
-      
-      /*
-       shadow_region_t *ent = Malloc(sizeof(shadow_region_t));;
-       init_shadow_region_physical(ent,0,0x100000,GUEST_REGION_PHYSICAL_MEMORY,
-       0x100000, HOST_REGION_PHYSICAL_MEMORY);
-       add_shadow_region(&(vm_info.mem_map),ent);
-      */
-
-      add_shadow_region_passthrough(&vm_info, 0x0, 0x100000, 0x100000);
-
-      v3_hook_io_port(&vm_info, 0x61, &IO_Read, &IO_Write, NULL);
-      v3_hook_io_port(&vm_info, 0x05, &IO_Read, &IO_Write_to_Serial, NULL);
-      
-      /*
-       vm_info.cr0 = 0;
-       vm_info.cs.base=0xf000;
-       vm_info.cs.limit=0xffff;
-      */
-      //vm_info.rip = 0xfff0;
-
-      vm_info.rip = 0;
-      vm_info.vm_regs.rsp = 0x0;
-    } else {
-   
+       if (v3_add_shadow_mem(vm, V3_MEM_CORE_ANY, start_addr, end_addr, host_addr) == -1) {
+           PrintError("Could not map memory region: %p-%p => %p\n", 
+                      (void *)start_addr, (void *)end_addr, (void *)host_addr);
+           return -1;
+       }
+
+       mem_region = v3_cfg_next_branch(mem_region);
     }
 
-#endif
+    return 0;
+}
+
+
+
+
+static int setup_devices(struct v3_vm_info * vm, v3_cfg_tree_t * cfg) {
+    v3_cfg_tree_t * device = v3_cfg_subtree(v3_cfg_subtree(cfg, "devices"), "device");
+
+    
+    while (device) {
+       char * dev_class = v3_cfg_val(device, "class");
+
+       V3_Print("configuring device %s\n", dev_class);
+
+       if (v3_create_device(vm, dev_class, device) == -1) {
+           PrintError("Error creating device %s\n", dev_class);
+           return -1;
+       }
+       
+       device = v3_cfg_next_branch(device);
+    }
+
+    v3_print_dev_mgr(vm);
+
+    return 0;
+}
+
+
+
+