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.


Completed preliminary host and guest state structure for VMLAUNCH. Added assembly...
[palacios.git] / palacios / src / palacios / vmx.c
index 6195141..6c2939d 100644 (file)
@@ -24,6 +24,9 @@
 #include <palacios/vmcs.h>
 #include <palacios/vmm.h>
 #include <palacios/vmx_lowlevel.h>
+#include <palacios/vmm_lowlevel.h>
+#include <palacios/vmm_config.h>
+#include <palacios/vmm_ctrl_regs.h>
 
 
 // 
@@ -32,7 +35,7 @@
 //
 //
 
-
+#if 0
 
 #include <palacios/vmm_util.h>
 #include <palacios/vmm_string.h>
@@ -75,8 +78,6 @@ uint_t myregs = 0;
 
 
 
-static struct vmcs_data* vmxon_ptr;
-
 
 
 extern uint_t VMCS_LAUNCH();
@@ -121,61 +122,6 @@ void DecodeCurrentInstruction(struct VM *vm, struct Instruction *inst)
 }
 
 
-static void setup_v8086_mode_for_boot(struct guest_info* vm_info)
-{
-
-    ((struct vmx_data*)vm_info->vmm_data)->state = VMXASSIST_V8086_BIOS;
-    ((struct rflags)info->ctrl_regs.rflags).vm = 1;
-    ((struct rflags)info->ctrl_regs.rflags).iopl = 3;
-
-
-    vm_info->rip = 0xfff0;
-
-    vm_info->segments.cs.selector = 0xf000;
-    vm_info->segments.cs.base = 0xf000<<4;
-    vm_info->segments.cs.limit = 0xffff;
-    vm_info->segments.cs.type = 3;
-    vm_info->segments.cs.system = 1;
-    vm_info->segments.cs.dpl = 3;
-    vm_info->segments.cs.present = 1;
-    vm_info->segments.cs.granularity = 0;
-
-    vm_info->segments.ss.selector = 0x0000;
-    vm_info->segments.ss.base = 0x0000<<4;
-    vm_info->segments.ss.limit = 0xffff;
-    vm_info->segments.ss.type = 3;
-    vm_info->segments.ss.system = 1;
-    vm_info->segments.ss.dpl = 3;
-    vm_info->segments.ss.present = 1;
-    vm_info->segments.ss.granularity = 0;
-
-    vm_info->segments.es.selector = 0x0000;
-    vm_info->segments.es.base = 0x0000<<4;
-    vm_info->segments.es.limit = 0xffff;
-    vm_info->segments.es.type = 3;
-    vm_info->segments.es.system = 1;
-    vm_info->segments.es.dpl = 3;
-    vm_info->segments.es.present = 1;
-    vm_info->segments.es.granularity = 0;
-
-    vm_info->segments.fs.selector = 0x0000;
-    vm_info->segments.fs.base = 0x0000<<4;
-    vm_info->segments.fs.limit = 0xffff;
-    vm_info->segments.fs.type = 3;
-    vm_info->segments.fs.system = 1;
-    vm_info->segments.fs.dpl = 3;
-    vm_info->segments.fs.present = 1;
-    vm_info->segments.fs.granularity = 0;
-
-    vm_info->segments.gs.selector = 0x0000;
-    vm_info->segments.gs.base = 0x0000<<4;
-    vm_info->segments.gs.limit = 0xffff;
-    vm_info->segments.gs.type = 3;
-    vm_info->segments.gs.system = 1;
-    vm_info->segments.gs.dpl = 3;
-    vm_info->segments.gs.present = 1;
-    vm_info->segments.gs.granularity = 0;
-}
 
 static void ConfigureExits(struct VM *vm)
 {
@@ -342,74 +288,49 @@ int VMLaunch(struct VMDescriptor *vm)
 //
 //
 
+#endif
+
 static int update_vmcs_host_state(struct guest_info * info) {
     addr_t tmp;
-
-    struct {
-       uint16 limit;
-       addr_t base;
-    } __attribute__((packed)) tmp_seg;
-
-
+    struct vmx_data * arch_data = (struct vmx_data *)(info->vmm_data);
     struct v3_msr tmp_msr;
 
-    __asm__ __volatile__ ( "movq    %%cr0, %1; "               
+    __asm__ __volatile__ ( "movq    %%cr0, %0; "               
                           : "=q"(tmp)
                           :
     );
-    vmcs_write(HOST_CR0, tmp);
+    vmcs_write(VMCS_HOST_CR0, tmp);
 
 
     __asm__ __volatile__ ( "movq %%cr3, %0; "          
                           : "=q"(tmp)
                           :
     );
-    vmcs_write(HOST_CR3, tmp);
+    vmcs_write(VMCS_HOST_CR3, tmp);
 
 
     __asm__ __volatile__ ( "movq %%cr4, %0; "          
                           : "=q"(tmp)
                           :
     );
-    vmcs_write(HOST_CR4, tmp);
-
-
-
-
-    __asm__ __volatile__ ("sgdt (%0); "
-                         : 
-                         :"q"(&tmp_seg)
-                         : "memory"
-                         );
-    vmcs_write(HOST_GDTR_BASE, tmp_seg.base);
+    vmcs_write(VMCS_HOST_CR4, tmp);
 
 
-    __asm__ __volatile__ ("sidt (%0); "
-                         : 
-                         :"q"(&tmp_seg)
-                         : "memory"
-                 );
-    vmcs_write(HOST_IDTR_BASE, tmp_seg.base);
-
-
-    __asm__ __volatile__ ("str (%0); "
-                         : 
-                         :"q"(&tmp_seg)
-                         : "memory"
-                         );
-    vmcs_write(HOST_TR_BASE, tmp_seg.base);
 
+    vmcs_write(VMCS_HOST_GDTR_BASE, arch_data->host_state.gdtr.base);
+    vmcs_write(VMCS_HOST_IDTR_BASE, arch_data->host_state.idtr.base);
+    vmcs_write(VMCS_HOST_TR_BASE, arch_data->host_state.tr.base);
 
 #define FS_BASE_MSR 0xc0000100
 #define GS_BASE_MSR 0xc0000101
 
     // FS.BASE MSR
     v3_get_msr(FS_BASE_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
-    vmcs_write(HOST_FS_BASE, tmp_msr.value);    
+    vmcs_write(VMCS_HOST_FS_BASE, tmp_msr.value);    
 
     // GS.BASE MSR
     v3_get_msr(GS_BASE_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
-    vmcs_write(HOST_GS_BASE, tmp_msr.value);    
+    vmcs_write(VMCS_HOST_GS_BASE, tmp_msr.value);    
 
 
 
@@ -431,6 +352,12 @@ static int update_vmcs_host_state(struct guest_info * info) {
     );
     vmcs_write(VMCS_HOST_DS_SELECTOR, tmp);
 
+    __asm__ __volatile__ ( "movq %%es, %0; "           
+                          : "=q"(tmp)
+                          :
+    );
+    vmcs_write(VMCS_HOST_ES_SELECTOR, tmp);
+
     __asm__ __volatile__ ( "movq %%fs, %0; "           
                           : "=q"(tmp)
                           :
@@ -443,11 +370,7 @@ static int update_vmcs_host_state(struct guest_info * info) {
     );
     vmcs_write(VMCS_HOST_GS_SELECTOR, tmp);
 
-    __asm__ __volatile__ ( "movq %%tr, %0; "           
-                          : "=q"(tmp)
-                          :
-    );
-    vmcs_write(VMCS_HOST_TR_SELECTOR, tmp);
+    vmcs_write(VMCS_HOST_TR_SELECTOR, arch_data->host_state.tr.selector);
 
 
 #define SYSENTER_CS_MSR 0x00000174
@@ -456,28 +379,30 @@ static int update_vmcs_host_state(struct guest_info * info) {
 
    // SYSENTER CS MSR
     v3_get_msr(SYSENTER_CS_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
-    vmcs_write(HOST_IA32_SYSENTER_CS, tmp_msr.value);    
+    vmcs_write(VMCS_HOST_SYSENTER_CS, tmp_msr.value);
 
     // SYSENTER_ESP MSR
     v3_get_msr(SYSENTER_ESP_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
-    vmcs_write(HOST_IA32_SYSENTER_ESP, tmp_msr.value);    
+    vmcs_write(VMCS_HOST_SYSENTER_ESP, tmp_msr.value);
 
     // SYSENTER_EIP MSR
     v3_get_msr(SYSENTER_EIP_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
-    vmcs_write(HOST_IA32_SYSENTER_EIP, tmp_msr.value);    
+    vmcs_write(VMCS_HOST_SYSENTER_EIP, tmp_msr.value);
 
 
-    // RIP
-    // RSP
 
     return 0;
-
 }
 
 
 
 
+static addr_t vmxon_ptr_phys;
+extern int v3_vmx_exit_handler();
+extern int v3_vmx_vmlaunch();
+
+
+#if 0
 // For the 32 bit reserved bit fields 
 // MB1s are in the low 32 bits, MBZs are in the high 32 bits of the MSR
 static uint32_t sanitize_bits1(uint32_t msr_num, uint32_t val) {
@@ -487,7 +412,7 @@ static uint32_t sanitize_bits1(uint32_t msr_num, uint32_t val) {
 
     v3_get_msr(msr_num, &mask_msr.hi, &mask_msr.lo);
 
-    PrintDebug("MSR %x = %x : %x \n", msr_num, msr.hi, msr.lo);
+    PrintDebug("MSR %x = %x : %x \n", msr_num, mask_msr.hi, mask_msr.lo);
 
     val &= mask_msr.lo;
     val &= mask_msr.hi;
@@ -496,6 +421,7 @@ static uint32_t sanitize_bits1(uint32_t msr_num, uint32_t val) {
 }
 
 
+
 static addr_t sanitize_bits2(uint32_t msr_num0, uint32_t msr_num1, addr_t val) {
     v3_msr_t msr0, msr1;
     addr_t msr0_val, msr1_val;
@@ -509,7 +435,7 @@ static addr_t sanitize_bits2(uint32_t msr_num0, uint32_t msr_num1, addr_t val) {
     msr0_val = msr0.value;
     msr1_val = msr1.value;
 
-    PrintDebug("MSR %x = %p, %x = %p \n", msr_num0, msr0_val, msr_num1, msr1_val);
+    PrintDebug("MSR %x = %p, %x = %p \n", msr_num0, (void*)msr0_val, msr_num1, (void*)msr1_val);
 
     val &= msr0_val;
     val &= msr1_val;
@@ -517,71 +443,362 @@ static addr_t sanitize_bits2(uint32_t msr_num0, uint32_t msr_num1, addr_t val) {
     return val;
 }
 
+static int setup_base_host_state() {
+    
+
+
+    //   vmwrite(HOST_IDTR_BASE, 
+
+
+}
+
+
+#endif
+
+static int inline check_vmcs_write(vmcs_field_t field, addr_t val)
+{
+    int ret = 0;
+    ret = vmcs_write(field,val);
+
+    if (ret != VMX_SUCCESS) {
+        PrintError("VMWRITE error on %s!: %d\n", v3_vmcs_field_to_str(field), ret);
+        return 1;
+    }
+
+    return 0;
+}
+
+
+static void inline translate_segment_access(struct v3_segment * v3_seg,  
+                                           struct vmcs_segment_access * access)
+{
+    access->type = v3_seg->type;
+    access->desc_type = v3_seg->system;
+    access->dpl = v3_seg->dpl;
+    access->present = v3_seg->present;
+    access->avail = v3_seg->avail;
+    access->long_mode = v3_seg->long_mode;
+    access->db = v3_seg->db;
+    access->granularity = v3_seg->granularity;
+}
+
+static int inline vmcs_write_guest_segments(struct guest_info* info)
+{
+    int ret = 0;
+    struct vmcs_segment_access access;
+
+    memset(&access, 0, sizeof(access));
 
+    /* CS Segment */
+    translate_segment_access(&(info->segments.cs), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_CS_BASE, info->segments.cs.base);
+    ret &= check_vmcs_write(VMCS_GUEST_CS_SELECTOR, info->segments.cs.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_CS_LIMIT, info->segments.cs.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_CS_ACCESS, access.value);
+
+    /* SS Segment */
+    translate_segment_access(&(info->segments.ss), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_SS_BASE, info->segments.ss.base);
+    ret &= check_vmcs_write(VMCS_GUEST_SS_SELECTOR, info->segments.ss.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_SS_LIMIT, info->segments.ss.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_SS_ACCESS, access.value);
+
+    /* DS Segment */
+    translate_segment_access(&(info->segments.ds), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_DS_BASE, info->segments.ds.base);
+    ret &= check_vmcs_write(VMCS_GUEST_DS_SELECTOR, info->segments.ds.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_DS_LIMIT, info->segments.ds.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_DS_ACCESS, access.value);
+
+
+    /* ES Segment */
+    translate_segment_access(&(info->segments.es), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_ES_BASE, info->segments.es.base);
+    ret &= check_vmcs_write(VMCS_GUEST_ES_SELECTOR, info->segments.es.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_ES_LIMIT, info->segments.es.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_ES_ACCESS, access.value);
+
+    /* FS Segment */
+    translate_segment_access(&(info->segments.fs), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_FS_BASE, info->segments.fs.base);
+    ret &= check_vmcs_write(VMCS_GUEST_FS_SELECTOR, info->segments.fs.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_FS_LIMIT, info->segments.fs.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_FS_ACCESS, access.value);
+
+    /* GS Segment */
+    translate_segment_access(&(info->segments.gs), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_GS_BASE, info->segments.gs.base);
+    ret &= check_vmcs_write(VMCS_GUEST_GS_SELECTOR, info->segments.gs.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_GS_LIMIT, info->segments.gs.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_GS_ACCESS, access.value);
+
+    /* LDTR segment */
+    translate_segment_access(&(info->segments.ldtr), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_LDTR_BASE, info->segments.ldtr.base);
+    ret &= check_vmcs_write(VMCS_GUEST_LDTR_SELECTOR, info->segments.ldtr.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_LDTR_LIMIT, info->segments.ldtr.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_LDTR_ACCESS, access.value);
+
+    /* TR Segment */
+    translate_segment_access(&(info->segments.tr), &access);
+
+    ret &= check_vmcs_write(VMCS_GUEST_TR_BASE, info->segments.tr.base);
+    ret &= check_vmcs_write(VMCS_GUEST_TR_SELECTOR, info->segments.ldtr.selector);
+    ret &= check_vmcs_write(VMCS_GUEST_TR_LIMIT, info->segments.tr.limit);
+    ret &= check_vmcs_write(VMCS_GUEST_TR_ACCESS, access.value);
+
+    /* GDTR Segment */
+
+    ret &= check_vmcs_write(VMCS_GUEST_GDTR_BASE, info->segments.gdtr.base);
+    ret &= check_vmcs_write(VMCS_GUEST_GDTR_LIMIT, info->segments.gdtr.limit);
+
+    /* IDTR Segment*/
+    ret &= check_vmcs_write(VMCS_GUEST_IDTR_BASE, info->segments.idtr.base);
+    ret &= check_vmcs_write(VMCS_GUEST_IDTR_LIMIT, info->segments.idtr.limit);
+
+    return ret;
+}
 
-static vmcs_data* allocate_vmcs() {
+static void setup_v8086_mode_for_boot(struct guest_info * vm_info)
+{
+
+    ((struct vmx_data *)vm_info->vmm_data)->state = VMXASSIST_V8086_BIOS;
+    ((struct rflags *)&(vm_info->ctrl_regs.rflags))->vm = 1;
+    ((struct rflags *)&(vm_info->ctrl_regs.rflags))->iopl = 3;
+
+   
+    vm_info->rip = 0xfff0;
+
+    vm_info->segments.cs.selector = 0xf000;
+    vm_info->segments.cs.base = 0xf000 << 4;
+    vm_info->segments.cs.limit = 0xffff;
+    vm_info->segments.cs.type = 3;
+    vm_info->segments.cs.system = 1;
+    vm_info->segments.cs.dpl = 3;
+    vm_info->segments.cs.present = 1;
+    vm_info->segments.cs.granularity = 0;
+
+    int i = 0;
+    struct v3_segment * seg_ptr = (struct v3_segment *)&(vm_info->segments);
+
+    /* Set values for selectors ds through ss */
+    for(i = 1; i < 6 ; i++) {
+        seg_ptr[i].selector = 0x0000;
+        seg_ptr[i].base = 0x00000;
+        seg_ptr[i].type = 3;
+        seg_ptr[i].system = 1;
+        seg_ptr[i].dpl = 3;
+        seg_ptr[i].present = 1;
+        seg_ptr[i].granularity = 0;
+    }
+
+    for(i = 6; i < 10; i++) {
+        seg_ptr[i].base = 0x0;
+        seg_ptr[i].limit = 0xffff;
+    }
+
+    vm_info->segments.ldtr.selector = 0x0;
+    vm_info->segments.ldtr.type = 2;
+    vm_info->segments.ldtr.system = 0;
+    vm_info->segments.ldtr.present = 1;
+    vm_info->segments.ldtr.granularity = 0;
+
+    vm_info->segments.tr.selector = 0x0;
+    vm_info->segments.tr.type = 3;
+    vm_info->segments.tr.system = 0;
+    vm_info->segments.tr.present = 1;
+    vm_info->segments.tr.granularity = 0;
+}
+
+
+static addr_t allocate_vmcs() 
+{
     reg_ex_t msr;
-    vmcs_data* vmcs_page = (vmcs_data*)V3_VAddr(V3_AllocPages(1));
+    PrintDebug("Allocating page\n");
+    struct vmcs_data * vmcs_page = (struct vmcs_data *)V3_VAddr(V3_AllocPages(1));
+
 
     memset(vmcs_page, 0, 4096);
 
     v3_get_msr(VMX_BASIC_MSR, &(msr.e_reg.high), &(msr.e_reg.low));
     
-    vmcs_page->revision = ((struct vmx_basic_msr)msr).revision;
+    vmcs_page->revision = ((struct vmx_basic_msr*)&msr)->revision;
+    PrintDebug("VMX Revision: 0x%x\n",vmcs_page->revision);
 
-    return vmcs_page;
+    return (addr_t)V3_PAddr((void *)vmcs_page);
 }
 
 
 
-static void init_vmcs_bios(struct guest_info * vm_info) 
+static int init_vmcs_bios(struct guest_info * vm_info) 
 {
 
+    setup_v8086_mode_for_boot(vm_info);
 
-}
+    // TODO: Fix vmcs fields so they're 32-bit
+    struct vmx_data * vmx_data = (struct vmx_data *)vm_info->vmm_data;
+    int vmx_ret;
+
+    // Have to do a whole lot of flag setting here
+    PrintDebug("Clearing VMCS\n");
+    vmx_ret = vmcs_clear(vmx_data->vmcs_ptr_phys);
+
+    if (vmx_ret != VMX_SUCCESS) {
+        PrintError("VMCLEAR failed\n");
+        return -1;
+    }
+
+    PrintDebug("Loading VMCS\n");
+    vmx_ret = vmcs_load(vmx_data->vmcs_ptr_phys);
+
+    if (vmx_ret != VMX_SUCCESS) {
+        PrintError("VMPTRLD failed\n");
+        return -1;
+    }
+
+    struct v3_msr tmp_msr;
+
+    /* Write VMX Control Fields */
+    v3_get_msr(VMX_PINBASED_CTLS_MSR,&(tmp_msr.hi),&(tmp_msr.lo));
+    vmcs_write(VMCS_PIN_CTRLS, tmp_msr.lo);
+
+    v3_get_msr(VMX_PROCBASED_CTLS_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
+    vmcs_write(VMCS_PROC_CTRLS, tmp_msr.lo);
 
+    v3_get_msr(VMX_EXIT_CTLS_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
+    vmcs_write(VMCS_EXIT_CTRLS, tmp_msr.lo);
 
+    v3_get_msr(VMX_ENTRY_CTLS_MSR, &(tmp_msr.hi), &(tmp_msr.lo));
+    vmcs_write(VMCS_ENTRY_CTRLS, tmp_msr.lo);
+
+    /* Cache GDTR, IDTR, and TR in host struct */
+    struct {
+        uint16_t selector;
+        addr_t   base;
+    } __attribute__((packed)) tmp_seg;
+    
+    addr_t gdtr_base;
+
+    __asm__ __volatile__(
+                        "sgdt (%0);"
+                        :
+                        : "q"(&tmp_seg)
+                        : "memory"
+                        );
+    vmx_data->host_state.gdtr.base = gdtr_base = tmp_seg.base;
+
+    __asm__ __volatile__(
+                        "sidt (%0);"
+                        :
+                        : "q"(&tmp_seg)
+                        : "memory"
+                        );
+    vmx_data->host_state.idtr.base = tmp_seg.base;
+
+    __asm__ __volatile__(
+                        "str (%0);"
+                        :
+                        : "q"(&tmp_seg)
+                        : "memory"
+                        );
+    vmx_data->host_state.tr.selector = tmp_seg.selector;
+
+    struct tss_descriptor desc = ((struct tss_descriptor *)gdtr_base)[tmp_seg.selector];
+    
+    tmp_seg.base = (
+                   (desc.base1) |
+                   (desc.base2 << 16) |
+                   (desc.base3 << 24) |
+#ifdef __V3_64BIT__
+                   ((uint64_t)desc.base4 << 32)
+#else 
+                   (0)
+#endif
+                );
+
+    vmx_data->host_state.tr.base = tmp_seg.base;
+
+    update_vmcs_host_state(vm_info);
+    vmcs_write(VMCS_HOST_RIP, (addr_t)&v3_vmx_exit_handler);
+
+    // Setup guest state 
+    // TODO: This is not 32-bit safe!
+    vmx_ret &= check_vmcs_write(VMCS_GUEST_RIP, vm_info->rip);
+    vmx_ret &= check_vmcs_write(VMCS_GUEST_CR0, 0x60000010);
+
+    vmx_ret &= vmcs_write_guest_segments(vm_info);
+
+    vmx_ret &= check_vmcs_write(VMCS_GUEST_RFLAGS, vm_info->ctrl_regs.rflags);
+    vmx_ret &= check_vmcs_write(VMCS_LINK_PTR, 0xffffffffffffffff);
+
+    if (vmx_ret != 0) {
+       PrintError("Could not initialize VMCS segments\n");
+        return -1;
+    }
+
+    v3_print_vmcs_guest_state();
+    return 0;
+}
+
+int v3_vmx_handle_exit()
+{
+    PrintDebug("Exit taken!\n");
+    return 0;
+}
 
 static int init_vmx_guest(struct guest_info * info, struct v3_vm_config * config_ptr) {
+    PrintDebug("Entering init_vmx_guest\n");
     v3_pre_config_guest(info, config_ptr);
 
-    struct vmx_data* data;
+    struct vmx_data * data = NULL;
+
+    data = (struct vmx_data *)V3_Malloc(sizeof(struct vmx_data));
+
+    PrintDebug("vmx_data pointer: %p\n", (void *)data);
 
-    PrintDebug("Allocating vmx_data\n");
-    data = (struct vmx_data*)V3_Malloc(sizeof(struct vmx_data));
     PrintDebug("Allocating VMCS\n");
-    data->vmcs = allocate_vmcs();
+    data->vmcs_ptr_phys = allocate_vmcs();
+
+    PrintDebug("VMCS pointer: %p\n", (void *)(data->vmcs_ptr_phys));
 
-    info->vmm_data = (void*)data;
+    info->vmm_data = data;
 
     PrintDebug("Initializing VMCS (addr=%p)\n", info->vmm_data);
-    init_vmcs_bios((struct vmx_data*)(info->vmm_data), info);
 
-    v3_post_config_guest(info, config_ptr);
+    if (init_vmcs_bios(info) != 0) {
+       PrintError("Could not initialize VMCS BIOS\n");
+        return -1;
+    }
+
+    //v3_post_config_guest(info, config_ptr);
 
     return 0;
 }
 
 
+static int start_vmx_guest(struct guest_info* info) {
+    uint32_t error = 0;
+    int ret = 0;
 
+    PrintDebug("Attempting VMLAUNCH\n");
 
-static int start_vmx_guest(struct guest_info *info) {
-    struct vmx_data* vmx_data = (struct vmx_data*)info->vmm_data;
-    int vmx_ret;
+    ret = v3_vmx_vmlaunch();
 
-    // Have to do a whole lot of flag setting here
-    vmx_ret = vmcs_clear(vmx_data->vmcs);
-    if(vmx_ret != VMX_SUCCESS) {
-        PrintDebug("VMCS Clear failed\n");
-        return -1;
-    }
-    vmx_ret = vmcs_load(vmx_data->vmcs);
-    if(vmx_ret != VMX_SUCCESS) {
-        PrintDebug("Executing VMPTRLD\n");
-        return -1;
+    PrintDebug("Returned from VMLAUNCH\n");
+
+    vmcs_read(VMCS_INSTR_ERR, &error, 4);
+
+    if (ret != VMX_SUCCESS) {
+        PrintError("VMLAUNCH failed: %d\n", error);
     }
 
-    // Setup guest state
     return -1;
 }
 
@@ -591,21 +808,22 @@ static int start_vmx_guest(struct guest_info *info) {
 
 
 int v3_is_vmx_capable() {
-    uint_t ret;
     v3_msr_t feature_msr;
     addr_t eax = 0, ebx = 0, ecx = 0, edx = 0;
 
-    v3_cpuid(CPUID_FEATURE_IDS, &eax, &ebx, &ecx, &edx);
+    v3_cpuid(0x1, &eax, &ebx, &ecx, &edx);
+
+    PrintDebug("ECX: %p\n", (void*)ecx);
 
     if (ecx & CPUID_1_ECX_VTXFLAG) {
-       v3_get_msr(IA32_FEATURE_CONTROL_MSR, &(feature_msr.hi), &(feature_msr.lo));
+        v3_get_msr(VMX_FEATURE_CONTROL_MSR, &(feature_msr.hi), &(feature_msr.lo));
        
-       PrintTrace("MSRREGlow: 0x%.8x\n", feature_msr.lo);
+        PrintTrace("MSRREGlow: 0x%.8x\n", feature_msr.lo);
 
-       if ((feature_msr.lo & FEATURE_CONTROL_VALID) != FEATURE_CONTROL_VALID) {
-           PrintDebug("VMX is locked -- enable in the BIOS\n");
-           return 0;
-       }
+        if ((feature_msr.lo & FEATURE_CONTROL_VALID) != FEATURE_CONTROL_VALID) {
+            PrintDebug("VMX is locked -- enable in the BIOS\n");
+            return 0;
+        }
 
     } else {
         PrintDebug("VMX not supported on this cpu\n");
@@ -621,50 +839,49 @@ static int has_vmx_nested_paging() {
 
 
 
-// We set up the global host state that is unlikely to change across processes here
-// Segment Descriptors mainly
-
-struct seg_descriptor {
-
-};
-
-
-static int setup_base_host_state() {
-    uint8_t gdt[10];
-    
-
-
-    //   vmwrite(HOST_IDTR_BASE, 
-
-
-}
-
-
-
 void v3_init_vmx(struct v3_ctrl_ops * vm_ops) {
-    v3_msr_t basic_msr;
+    extern v3_cpu_arch_t v3_cpu_type;
+
+    struct v3_msr tmp_msr;
+    uint64_t ret=0;
 
+    v3_get_msr(VMX_CR4_FIXED0_MSR,&(tmp_msr.hi),&(tmp_msr.lo));
     
     __asm__ __volatile__ (
-                         "movl %%cr4, %%ebx; "
-                         "orl  %%ebx, 0x00002000; "
-                         "movl %%ebx, %%cr4"
+                         "movq %%cr4, %%rbx;"
+                         "orq  $0x00002000, %%rbx;"
+                         "movq %%rbx, %0;"
+                         : "=m"(ret) 
+                         :
+                         : "%rbx"
                          );
 
-
-
+    if((~ret & tmp_msr.value) == 0) {
+        __asm__ __volatile__ (
+                             "movq %0, %%cr4;"
+                             :
+                             : "q"(ret)
+                             );
+    } else {
+        PrintError("Invalid CR4 Settings!\n");
+        return;
+    }
+      __asm__ __volatile__ (
+                           "movq %%cr0, %%rbx; "
+                           "orq  $0x00000020,%%rbx; "
+                           "movq %%rbx, %%cr0;"
+                           :
+                           :
+                           : "%rbx"
+                           );
+      //
     // Should check and return Error here.... 
-    __asm__ __volatile__ (
-                         "movl %%cr0, %%ebx; "
-                         "orl  %%ebx, 0x00000020; "
-                         "movl %%ebx, %%cr0"
-                         );
 
     // Setup VMXON Region
-    vmxon_ptr = allocate_vmcs();
-    PrintDebug("VMX revision: 0x%p\n", (void*)vmxon_ptr);
+    vmxon_ptr_phys = allocate_vmcs();
+    PrintDebug("VMXON pointer: 0x%p\n", (void*)vmxon_ptr_phys);
 
-    if (v3_enable_vmx(vmxon_ptr) == 0) {
+    if (v3_enable_vmx(vmxon_ptr_phys) == VMX_SUCCESS) {
         PrintDebug("VMX Enabled\n");
     } else {
         PrintError("VMX initialization failure\n");
@@ -679,8 +896,8 @@ void v3_init_vmx(struct v3_ctrl_ops * vm_ops) {
     }
 
     // Setup the VMX specific vmm operations
-    vmm_ops->init_guest = &init_vmx_guest;
-    vmm_ops->start_guest = &start_vmx_guest;
-    vmm_ops->has_nested_paging = &has_vmx_nested_paging;
+    vm_ops->init_guest = &init_vmx_guest;
+    vm_ops->start_guest = &start_vmx_guest;
+    vm_ops->has_nested_paging = &has_vmx_nested_paging;
 
 }