-/* (c) 2008, Jack Lange <jarusl@cs.northwestern.edu> */
-/* (c) 2008, The V3VEE Project <http://www.v3vee.org> */
-
+/*
+ * 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_mem.h>
#include <palacios/vmm.h>
#include <palacios/vmm_decoder.h>
#include <palacios/vm_guest_mem.h>
#include <palacios/vmm_ctrl_regs.h>
+#include <palacios/vmm_direct_paging.h>
+#include <palacios/svm.h>
-
-
-/* Segmentation is a problem here...
- *
- * When we get a memory operand, presumably we use the default segment (which is?)
- * unless an alternate segment was specfied in the prefix...
- */
-
-
-#ifndef DEBUG_CTRL_REGS
+#ifndef V3_CONFIG_DEBUG_CTRL_REGS
#undef PrintDebug
#define PrintDebug(fmt, args...)
#endif
+static int handle_lmsw(struct guest_info * info, struct x86_instr * dec_instr);
+static int handle_clts(struct guest_info * info, struct x86_instr * dec_instr);
+static int handle_mov_to_cr0(struct guest_info * info, struct x86_instr * dec_instr);
+
+
// First Attempt = 494 lines
// current = 106 lines
-int handle_cr0_write(struct guest_info * info) {
- char instr[15];
- int ret;
- struct x86_instr dec_instr;
-
- if (info->mem_mode == PHYSICAL_MEM) {
- ret = read_guest_pa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- } else {
- ret = read_guest_va_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- }
-
- /* The IFetch will already have faulted in the necessary bytes for the full instruction
- if (ret != 15) {
- // I think we should inject a GPF into the guest
- PrintError("Could not read instruction (ret=%d)\n", ret);
- return -1;
- }
- */
-
- if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
- PrintError("Could not decode instruction\n");
- return -1;
- }
-
-
- if (opcode_cmp(V3_OPCODE_LMSW, (const uchar_t *)(dec_instr.opcode)) == 0) {
- struct cr0_real *real_cr0 = (struct cr0_real*)&(info->ctrl_regs.cr0);
- struct cr0_real *new_cr0 = (struct cr0_real *)(dec_instr.src_operand.operand);
- uchar_t new_cr0_val;
+int v3_handle_cr0_write(struct guest_info * info) {
+ uchar_t instr[15];
+ int ret;
+ struct x86_instr dec_instr;
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ }
- PrintDebug("LMSW\n");
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
+ }
+
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
- new_cr0_val = (*(char*)(new_cr0)) & 0x0f;
- PrintDebug("OperandVal = %x\n", new_cr0_val);
+ if (dec_instr.op_type == V3_OP_LMSW) {
+ if (handle_lmsw(info, &dec_instr) == -1) {
+ return -1;
+ }
+ } else if (dec_instr.op_type == V3_OP_MOV2CR) {
+ if (handle_mov_to_cr0(info, &dec_instr) == -1) {
+ return -1;
+ }
+ } else if (dec_instr.op_type == V3_OP_CLTS) {
+ if (handle_clts(info, &dec_instr) == -1) {
+ return -1;
+ }
+ } else {
+ PrintError(info->vm_info, info, "Unhandled opcode in handle_cr0_write\n");
+ return -1;
+ }
+
+ info->rip += dec_instr.instr_length;
+
+ return 0;
+}
- PrintDebug("Old CR0=%x\n", *real_cr0);
- *(uchar_t*)real_cr0 &= 0xf0;
- *(uchar_t*)real_cr0 |= new_cr0_val;
- PrintDebug("New CR0=%x\n", *real_cr0);
-
- if (info->shdw_pg_mode == SHADOW_PAGING) {
- struct cr0_real * shadow_cr0 = (struct cr0_real*)&(info->shdw_pg_state.guest_cr0);
-
- PrintDebug(" Old Shadow CR0=%x\n", *shadow_cr0);
- *(uchar_t*)shadow_cr0 &= 0xf0;
- *(uchar_t*)shadow_cr0 |= new_cr0_val;
- PrintDebug("New Shadow CR0=%x\n", *shadow_cr0);
+
+
+// The CR0 register only has flags in the low 32 bits
+// The hardware does a format check to make sure the high bits are zero
+// Because of this we can ignore the high 32 bits here
+static int handle_mov_to_cr0(struct guest_info * info, struct x86_instr * dec_instr) {
+ // 32 bit registers
+ struct cr0_32 * shadow_cr0 = (struct cr0_32 *)&(info->ctrl_regs.cr0);
+ struct cr0_32 * new_cr0 = (struct cr0_32 *)(dec_instr->src_operand.operand);
+ struct cr0_32 * guest_cr0 = (struct cr0_32 *)&(info->shdw_pg_state.guest_cr0);
+ uint_t paging_transition = 0;
+
+ PrintDebug(info->vm_info, info, "MOV2CR0 (MODE=%s)\n", v3_cpu_mode_to_str(info->cpu_mode));
+
+ PrintDebug(info->vm_info, info, "OperandVal = %x, length=%d\n", *(uint_t *)new_cr0, dec_instr->src_operand.size);
+
+ PrintDebug(info->vm_info, info, "Old CR0=%x\n", *(uint_t *)shadow_cr0);
+ PrintDebug(info->vm_info, info, "Old Guest CR0=%x\n", *(uint_t *)guest_cr0);
+
+
+ // We detect if this is a paging transition
+ if (guest_cr0->pg != new_cr0->pg) {
+ paging_transition = 1;
+ }
+
+ // Guest always sees the value they wrote
+ *guest_cr0 = *new_cr0;
+
+ // This value must always be set to 1
+ guest_cr0->et = 1;
+
+ // Set the shadow register to catch non-virtualized flags
+ *shadow_cr0 = *guest_cr0;
+
+ // Paging is always enabled
+ shadow_cr0->pg = 1;
+
+ if (guest_cr0->pg == 0) {
+ // If paging is not enabled by the guest, then we always enable write-protect to catch memory hooks
+ shadow_cr0->wp = 1;
}
- } else if (opcode_cmp(V3_OPCODE_MOV2CR, (const uchar_t *)(dec_instr.opcode)) == 0) {
- PrintDebug("MOV2CR0\n");
+
+ // Was there a paging transition
+ // Meaning we need to change the page tables
+ if (paging_transition) {
+ if (v3_get_vm_mem_mode(info) == VIRTUAL_MEM) {
+
+ struct efer_64 * guest_efer = (struct efer_64 *)&(info->shdw_pg_state.guest_efer);
+ struct efer_64 * shadow_efer = (struct efer_64 *)&(info->ctrl_regs.efer);
+
+ // Check long mode LME to set LME
+ if (guest_efer->lme == 1) {
+ PrintDebug(info->vm_info, info, "Enabing Long Mode\n");
+ guest_efer->lma = 1;
+
+ shadow_efer->lma = 1;
+ shadow_efer->lme = 1;
+
+ PrintDebug(info->vm_info, info, "New EFER %p\n", (void *)*(addr_t *)(shadow_efer));
+ }
+
+ PrintDebug(info->vm_info, info, "Activating Shadow Page Tables\n");
+
+ if (v3_activate_shadow_pt(info) == -1) {
+ PrintError(info->vm_info, info, "Failed to activate shadow page tables\n");
+ return -1;
+ }
+ } else {
- if (info->cpu_mode == LONG) {
- // 64 bit registers
- } else {
- // 32 bit registers
- struct cr0_32 *real_cr0 = (struct cr0_32*)&(info->ctrl_regs.cr0);
- struct cr0_32 *new_cr0= (struct cr0_32 *)(dec_instr.src_operand.operand);
+ shadow_cr0->wp = 1;
+
+ if (v3_activate_passthrough_pt(info) == -1) {
+ PrintError(info->vm_info, info, "Failed to activate passthrough page tables\n");
+ return -1;
+ }
+ }
+ }
+
+
+ PrintDebug(info->vm_info, info, "New Guest CR0=%x\n",*(uint_t *)guest_cr0);
+ PrintDebug(info->vm_info, info, "New CR0=%x\n", *(uint_t *)shadow_cr0);
+
+ return 0;
+}
- PrintDebug("OperandVal = %x, length=%d\n", *new_cr0, dec_instr.src_operand.size);
- PrintDebug("Old CR0=%x\n", *real_cr0);
- *real_cr0 = *new_cr0;
-
- if (info->shdw_pg_mode == SHADOW_PAGING) {
- struct cr0_32 * shadow_cr0 = (struct cr0_32 *)&(info->shdw_pg_state.guest_cr0);
-
- PrintDebug("Old Shadow CR0=%x\n", *shadow_cr0);
-
- real_cr0->et = 1;
-
- *shadow_cr0 = *new_cr0;
- shadow_cr0->et = 1;
-
- if (get_mem_mode(info) == VIRTUAL_MEM) {
- struct cr3_32 * shadow_cr3 = (struct cr3_32 *)&(info->shdw_pg_state.shadow_cr3);
-
- info->ctrl_regs.cr3 = *(addr_t*)shadow_cr3;
- } else {
- info->ctrl_regs.cr3 = *(addr_t*)&(info->direct_map_pt);
- real_cr0->pg = 1;
- }
-
- PrintDebug("New Shadow CR0=%x\n",*shadow_cr0);
- }
- PrintDebug("New CR0=%x\n", *real_cr0);
- }
-
- } else if (opcode_cmp(V3_OPCODE_CLTS, (const uchar_t *)(dec_instr.opcode)) == 0) {
+static int handle_clts(struct guest_info * info, struct x86_instr * dec_instr) {
// CLTS
- struct cr0_32 *real_cr0 = (struct cr0_32*)&(info->ctrl_regs.cr0);
-
+ struct cr0_32 * real_cr0 = (struct cr0_32*)&(info->ctrl_regs.cr0);
+
real_cr0->ts = 0;
-
+
if (info->shdw_pg_mode == SHADOW_PAGING) {
- struct cr0_32 * shadow_cr0 = (struct cr0_32 *)&(info->shdw_pg_state.guest_cr0);
- shadow_cr0->ts = 0;
+ struct cr0_32 * guest_cr0 = (struct cr0_32 *)&(info->shdw_pg_state.guest_cr0);
+ guest_cr0->ts = 0;
}
- } else {
- PrintError("Unhandled opcode in handle_cr0_write\n");
- return -1;
- }
+ return 0;
+}
- info->rip += dec_instr.instr_length;
- return 0;
+static int handle_lmsw(struct guest_info * info, struct x86_instr * dec_instr) {
+ struct cr0_real * real_cr0 = (struct cr0_real *)&(info->ctrl_regs.cr0);
+ // XED is a mess, and basically reverses the operand order for an LMSW
+ struct cr0_real * new_cr0 = (struct cr0_real *)(dec_instr->dst_operand.operand);
+ uchar_t new_cr0_val;
+
+ PrintDebug(info->vm_info, info, "LMSW\n");
+
+ new_cr0_val = (*(char*)(new_cr0)) & 0x0f;
+
+ PrintDebug(info->vm_info, info, "OperandVal = %x\n", new_cr0_val);
+
+ // We can just copy the new value through
+ // we don't need to virtualize the lower 4 bits
+ PrintDebug(info->vm_info, info, "Old CR0=%x\n", *(uint_t *)real_cr0);
+ *(uchar_t*)real_cr0 &= 0xf0;
+ *(uchar_t*)real_cr0 |= new_cr0_val;
+ PrintDebug(info->vm_info, info, "New CR0=%x\n", *(uint_t *)real_cr0);
+
+
+ // If Shadow paging is enabled we push the changes to the virtualized copy of cr0
+ if (info->shdw_pg_mode == SHADOW_PAGING) {
+ struct cr0_real * guest_cr0 = (struct cr0_real*)&(info->shdw_pg_state.guest_cr0);
+
+ PrintDebug(info->vm_info, info, "Old Guest CR0=%x\n", *(uint_t *)guest_cr0);
+ *(uchar_t*)guest_cr0 &= 0xf0;
+ *(uchar_t*)guest_cr0 |= new_cr0_val;
+ PrintDebug(info->vm_info, info, "New Guest CR0=%x\n", *(uint_t *)guest_cr0);
+ }
+ return 0;
}
+
+
+
// First attempt = 253 lines
// current = 51 lines
-int handle_cr0_read(struct guest_info * info) {
- char instr[15];
- int ret;
- struct x86_instr dec_instr;
-
- if (info->mem_mode == PHYSICAL_MEM) {
- ret = read_guest_pa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- } else {
- ret = read_guest_va_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- }
-
- /* The IFetch will already have faulted in the necessary bytes for the full instruction
- if (ret != 15) {
- // I think we should inject a GPF into the guest
- PrintError("Could not read instruction (ret=%d)\n", ret);
- return -1;
- }
- */
-
- if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
- PrintError("Could not decode instruction\n");
- return -1;
- }
-
- if (opcode_cmp(V3_OPCODE_MOVCR2, (const uchar_t *)(dec_instr.opcode)) == 0) {
- struct cr0_32 * virt_cr0 = (struct cr0_32 *)(dec_instr.dst_operand.operand);
- struct cr0_32 * real_cr0 = (struct cr0_32 *)&(info->ctrl_regs.cr0);
-
- PrintDebug("MOVCR2\n");
- PrintDebug("CR0 at 0x%x\n", real_cr0);
+int v3_handle_cr0_read(struct guest_info * info) {
+ uchar_t instr[15];
+ int ret;
+ struct x86_instr dec_instr;
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ }
- if (info->shdw_pg_mode == SHADOW_PAGING) {
- *virt_cr0 = *(struct cr0_32 *)&(info->shdw_pg_state.guest_cr0);
- } else {
- *virt_cr0 = *real_cr0;
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
}
- PrintDebug("real CR0: %x\n", *(uint_t*)real_cr0);
- PrintDebug("returned CR0: %x\n", *(uint_t*)virt_cr0);
- } else if (opcode_cmp(V3_OPCODE_SMSW, (const uchar_t *)(dec_instr.opcode)) == 0) {
- struct cr0_real *real_cr0= (struct cr0_real*)&(info->ctrl_regs.cr0);
- struct cr0_real *virt_cr0 = (struct cr0_real *)(dec_instr.dst_operand.operand);
- char cr0_val = *(char*)real_cr0 & 0x0f;
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
- PrintDebug("SMSW\n");
+ if (dec_instr.op_type == V3_OP_MOVCR2) {
+ PrintDebug(info->vm_info, info, "MOVCR2 (mode=%s)\n", v3_cpu_mode_to_str(info->cpu_mode));
- PrintDebug("CR0 at 0x%x\n", real_cr0);
-
- *(char *)virt_cr0 &= 0xf0;
- *(char *)virt_cr0 |= cr0_val;
+ if ((v3_get_vm_cpu_mode(info) == LONG) ||
+ (v3_get_vm_cpu_mode(info) == LONG_32_COMPAT)) {
+ struct cr0_64 * dst_reg = (struct cr0_64 *)(dec_instr.dst_operand.operand);
+
+ if (info->shdw_pg_mode == SHADOW_PAGING) {
+ struct cr0_64 * guest_cr0 = (struct cr0_64 *)&(info->shdw_pg_state.guest_cr0);
+ *dst_reg = *guest_cr0;
+ } else {
+ struct cr0_64 * shadow_cr0 = (struct cr0_64 *)&(info->ctrl_regs.cr0);
+ *dst_reg = *shadow_cr0;
+ }
+
+ PrintDebug(info->vm_info, info, "returned CR0: %p\n", (void *)*(addr_t *)dst_reg);
+ } else {
+ struct cr0_32 * dst_reg = (struct cr0_32 *)(dec_instr.dst_operand.operand);
+
+ if (info->shdw_pg_mode == SHADOW_PAGING) {
+ struct cr0_32 * guest_cr0 = (struct cr0_32 *)&(info->shdw_pg_state.guest_cr0);
+ *dst_reg = *guest_cr0;
+ } else {
+ struct cr0_32 * shadow_cr0 = (struct cr0_32 *)&(info->ctrl_regs.cr0);
+ *dst_reg = *shadow_cr0;
+ }
+
+ PrintDebug(info->vm_info, info, "returned CR0: %x\n", *(uint_t*)dst_reg);
+ }
+
+ } else if (dec_instr.op_type == V3_OP_SMSW) {
+ struct cr0_real * shadow_cr0 = (struct cr0_real *)&(info->ctrl_regs.cr0);
+ struct cr0_real * dst_reg = (struct cr0_real *)(dec_instr.dst_operand.operand);
+ char cr0_val = *(char*)shadow_cr0 & 0x0f;
+
+ PrintDebug(info->vm_info, info, "SMSW\n");
+
+ // The lower 4 bits of the guest/shadow CR0 are mapped through
+ // We can treat nested and shadow paging the same here
+ *(char *)dst_reg &= 0xf0;
+ *(char *)dst_reg |= cr0_val;
+
+ } else {
+ PrintError(info->vm_info, info, "Unhandled opcode in handle_cr0_read\n");
+ return -1;
+ }
- } else {
- PrintError("Unhandled opcode in handle_cr0_read\n");
- return -1;
- }
-
- info->rip += dec_instr.instr_length;
+ info->rip += dec_instr.instr_length;
- return 0;
+ return 0;
}
+
// First Attempt = 256 lines
// current = 65 lines
-int handle_cr3_write(struct guest_info * info) {
- int ret;
- char instr[15];
- struct x86_instr dec_instr;
+int v3_handle_cr3_write(struct guest_info * info) {
+ int ret;
+ uchar_t instr[15];
+ struct x86_instr dec_instr;
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ }
+
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
+ }
+
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
+
+ if (dec_instr.op_type == V3_OP_MOV2CR) {
+ PrintDebug(info->vm_info, info, "MOV2CR3 (cpu_mode=%s)\n", v3_cpu_mode_to_str(info->cpu_mode));
+
+ if (info->shdw_pg_mode == SHADOW_PAGING) {
+ PrintDebug(info->vm_info, info, "Old Shadow CR3=%p; Old Guest CR3=%p\n",
+ (void *)(addr_t)(info->ctrl_regs.cr3),
+ (void*)(addr_t)(info->shdw_pg_state.guest_cr3));
+
+
+ // We update the guest CR3
+ if (info->cpu_mode == LONG) {
+ struct cr3_64 * new_cr3 = (struct cr3_64 *)(dec_instr.src_operand.operand);
+ struct cr3_64 * guest_cr3 = (struct cr3_64 *)&(info->shdw_pg_state.guest_cr3);
+ *guest_cr3 = *new_cr3;
+ } else {
+ struct cr3_32 * new_cr3 = (struct cr3_32 *)(dec_instr.src_operand.operand);
+ struct cr3_32 * guest_cr3 = (struct cr3_32 *)&(info->shdw_pg_state.guest_cr3);
+ *guest_cr3 = *new_cr3;
+ }
+
+
+ // If Paging is enabled in the guest then we need to change the shadow page tables
+ if (info->mem_mode == VIRTUAL_MEM) {
+ if (v3_activate_shadow_pt(info) == -1) {
+ PrintError(info->vm_info, info, "Failed to activate 32 bit shadow page table\n");
+ return -1;
+ }
+ }
+
+ PrintDebug(info->vm_info, info, "New Shadow CR3=%p; New Guest CR3=%p\n",
+ (void *)(addr_t)(info->ctrl_regs.cr3),
+ (void*)(addr_t)(info->shdw_pg_state.guest_cr3));
+
+ } else if (info->shdw_pg_mode == NESTED_PAGING) {
+
+ // This is just a passthrough operation which we probably don't need here
+ if (info->cpu_mode == LONG) {
+ struct cr3_64 * new_cr3 = (struct cr3_64 *)(dec_instr.src_operand.operand);
+ struct cr3_64 * guest_cr3 = (struct cr3_64 *)&(info->ctrl_regs.cr3);
+ *guest_cr3 = *new_cr3;
+ } else {
+ struct cr3_32 * new_cr3 = (struct cr3_32 *)(dec_instr.src_operand.operand);
+ struct cr3_32 * guest_cr3 = (struct cr3_32 *)&(info->ctrl_regs.cr3);
+ *guest_cr3 = *new_cr3;
+ }
+
+ }
+ } else {
+ PrintError(info->vm_info, info, "Unhandled opcode in handle_cr3_write\n");
+ return -1;
+ }
+
+ info->rip += dec_instr.instr_length;
+
+ return 0;
+}
- if (info->mem_mode == PHYSICAL_MEM) {
- ret = read_guest_pa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- } else {
- ret = read_guest_va_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- }
- /* The IFetch will already have faulted in the necessary bytes for the full instruction
- if (ret != 15) {
- // I think we should inject a GPF into the guest
- PrintError("Could not read instruction (ret=%d)\n", ret);
- return -1;
- }
- */
- if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
- PrintError("Could not decode instruction\n");
- return -1;
- }
+// first attempt = 156 lines
+// current = 36 lines
+int v3_handle_cr3_read(struct guest_info * info) {
+ uchar_t instr[15];
+ int ret;
+ struct x86_instr dec_instr;
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ }
+
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
+ }
+
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
+
+ if (dec_instr.op_type == V3_OP_MOVCR2) {
+ PrintDebug(info->vm_info, info, "MOVCR32 (mode=%s)\n", v3_cpu_mode_to_str(info->cpu_mode));
+
+ if (info->shdw_pg_mode == SHADOW_PAGING) {
+
+ if ((v3_get_vm_cpu_mode(info) == LONG) ||
+ (v3_get_vm_cpu_mode(info) == LONG_32_COMPAT)) {
+ struct cr3_64 * dst_reg = (struct cr3_64 *)(dec_instr.dst_operand.operand);
+ struct cr3_64 * guest_cr3 = (struct cr3_64 *)&(info->shdw_pg_state.guest_cr3);
+ *dst_reg = *guest_cr3;
+ } else {
+ struct cr3_32 * dst_reg = (struct cr3_32 *)(dec_instr.dst_operand.operand);
+ struct cr3_32 * guest_cr3 = (struct cr3_32 *)&(info->shdw_pg_state.guest_cr3);
+ *dst_reg = *guest_cr3;
+ }
+
+ } else if (info->shdw_pg_mode == NESTED_PAGING) {
+
+ // This is just a passthrough operation which we probably don't need here
+ if ((v3_get_vm_cpu_mode(info) == LONG) ||
+ (v3_get_vm_cpu_mode(info) == LONG_32_COMPAT)) {
+ struct cr3_64 * dst_reg = (struct cr3_64 *)(dec_instr.dst_operand.operand);
+ struct cr3_64 * guest_cr3 = (struct cr3_64 *)&(info->ctrl_regs.cr3);
+ *dst_reg = *guest_cr3;
+ } else {
+ struct cr3_32 * dst_reg = (struct cr3_32 *)(dec_instr.dst_operand.operand);
+ struct cr3_32 * guest_cr3 = (struct cr3_32 *)&(info->ctrl_regs.cr3);
+ *dst_reg = *guest_cr3;
+ }
+ }
+
+ } else {
+ PrintError(info->vm_info, info, "Unhandled opcode in handle_cr3_read\n");
+ return -1;
+ }
+
+ info->rip += dec_instr.instr_length;
+
+ return 0;
+}
- if (opcode_cmp(V3_OPCODE_MOV2CR, (const uchar_t *)(dec_instr.opcode)) == 0) {
- PrintDebug("MOV2CR3\n");
+//return guest cr4 - shadow PAE is always on
+int v3_handle_cr4_read(struct guest_info * info) {
+ uchar_t instr[15];
+ int ret;
+ struct x86_instr dec_instr;
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ }
+
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
+ }
+
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
+ if (dec_instr.op_type != V3_OP_MOVCR2) {
+ PrintError(info->vm_info, info, "Invalid opcode in read CR4\n");
+ return -1;
+ }
+
+ if (info->shdw_pg_mode == SHADOW_PAGING) {
+
+ if ((v3_get_vm_cpu_mode(info) == LONG) ||
+ (v3_get_vm_cpu_mode(info) == LONG_32_COMPAT)) {
+ struct cr4_64 * dst_reg = (struct cr4_64 *)(dec_instr.dst_operand.operand);
+ struct cr4_64 * guest_cr4 = (struct cr4_64 *)&(info->ctrl_regs.cr4);
+ *dst_reg = *guest_cr4;
+ }
+ else {
+ struct cr4_32 * dst_reg = (struct cr4_32 *)(dec_instr.dst_operand.operand);
+ struct cr4_32 * guest_cr4 = (struct cr4_32 *)&(info->shdw_pg_state.guest_cr4);
+ *dst_reg = *guest_cr4;
+ }
+
+ } else if (info->shdw_pg_mode == NESTED_PAGING) {
+
+
+ if ((v3_get_vm_cpu_mode(info) == LONG) ||
+ (v3_get_vm_cpu_mode(info) == LONG_32_COMPAT)) {
+ struct cr4_64 * dst_reg = (struct cr4_64 *)(dec_instr.dst_operand.operand);
+ struct cr4_64 * guest_cr4 = (struct cr4_64 *)&(info->ctrl_regs.cr4);
+ *dst_reg = *guest_cr4;
+ } else {
+ struct cr4_32 * dst_reg = (struct cr4_32 *)(dec_instr.dst_operand.operand);
+ struct cr4_32 * guest_cr4 = (struct cr4_32 *)&(info->ctrl_regs.cr4);
+ *dst_reg = *guest_cr4;
+ }
+ }
+
+ info->rip += dec_instr.instr_length;
+ return 0;
+}
- PrintDebug("CR3 at 0x%x\n", &(info->ctrl_regs.cr3));
- if (info->shdw_pg_mode == SHADOW_PAGING) {
- struct cr3_32 * new_cr3 = (struct cr3_32 *)(dec_instr.src_operand.operand);
- struct cr3_32 * guest_cr3 = (struct cr3_32 *)&(info->shdw_pg_state.guest_cr3);
- struct cr3_32 * shadow_cr3 = (struct cr3_32 *)&(info->shdw_pg_state.shadow_cr3);
- int cached = 0;
-
+int v3_handle_cr4_write(struct guest_info * info) {
+ uchar_t instr[15];
+ int ret;
+ int flush_tlb=0;
+ struct x86_instr dec_instr;
+ v3_cpu_mode_t cpu_mode = v3_get_vm_cpu_mode(info);
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ }
+
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
+ }
+
- PrintDebug("Old Shadow CR3=%x; Old Guest CR3=%x\n",
- *(uint_t*)shadow_cr3, *(uint_t*)guest_cr3);
-
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
+
+ if (dec_instr.op_type != V3_OP_MOV2CR) {
+ PrintError(info->vm_info, info, "Invalid opcode in write to CR4\n");
+ return -1;
+ }
+
+ // Check to see if we need to flush the tlb
+
- cached = cache_page_tables32(info, CR3_TO_PDE32(*(addr_t *)new_cr3));
+ if (v3_get_vm_mem_mode(info) == VIRTUAL_MEM) {
+ struct cr4_32 * new_cr4 = (struct cr4_32 *)(dec_instr.src_operand.operand);
+ struct cr4_32 * cr4 = (struct cr4_32 *)&(info->shdw_pg_state.guest_cr4);
+
+ // if pse, pge, or pae have changed while PG (in any mode) is on
+ // the side effect is a TLB flush, which means we need to
+ // toss the current shadow page tables too
+ //
+ //
+ // TODO - PAE FLAG needs to be special cased
+ if ((cr4->pse != new_cr4->pse) ||
+ (cr4->pge != new_cr4->pge) ||
+ (cr4->pae != new_cr4->pae)) {
+ PrintDebug(info->vm_info, info, "Handling PSE/PGE/PAE -> TLBFlush case, flag set\n");
+ flush_tlb = 1;
+
+ }
+ }
+
- if (cached == -1) {
- PrintError("CR3 Cache failed\n");
- return -1;
- } else if (cached == 0) {
- addr_t shadow_pt;
+ if ((cpu_mode == PROTECTED) || (cpu_mode == PROTECTED_PAE)) {
+ struct cr4_32 * new_cr4 = (struct cr4_32 *)(dec_instr.src_operand.operand);
+ struct cr4_32 * shadow_cr4 = (struct cr4_32 *)&(info->ctrl_regs.cr4);
+ struct cr4_32 * guest_cr4 = (struct cr4_32 *)&(info->shdw_pg_state.guest_cr4);
+ PrintDebug(info->vm_info, info, "OperandVal = %x, length = %d\n", *(uint_t *)new_cr4, dec_instr.src_operand.size);
+ PrintDebug(info->vm_info, info, "Old guest CR4=%x\n", *(uint_t *)guest_cr4);
- PrintDebug("New CR3 is different - flushing shadow page table\n");
+ if ((info->shdw_pg_mode == SHADOW_PAGING)) {
+ if (v3_get_vm_mem_mode(info) == PHYSICAL_MEM) {
+
+ if ((guest_cr4->pae == 0) && (new_cr4->pae == 1)) {
+ PrintDebug(info->vm_info, info, "Creating PAE passthrough tables\n");
+
+ // create 32 bit PAE direct map page table
+ if (v3_reset_passthrough_pts(info) == -1) {
+ PrintError(info->vm_info, info, "Could not create 32 bit PAE passthrough pages tables\n");
+ return -1;
+ }
+
+ // reset cr3 to new page tables
+ info->ctrl_regs.cr3 = *(addr_t*)&(info->direct_map_pt);
+
+ } else if ((guest_cr4->pae == 1) && (new_cr4->pae == 0)) {
+ // Create passthrough standard 32bit pagetables
+ PrintError(info->vm_info, info, "Switching From PAE to Protected mode not supported\n");
+ return -1;
+ }
+ }
+ }
- delete_page_tables_pde32((pde32_t *)CR3_TO_PDE32(*(uint_t*)shadow_cr3));
+ *guest_cr4 = *new_cr4;
+ *shadow_cr4 = *guest_cr4;
+ shadow_cr4->pae = 1; // always on for the shadow pager
+ PrintDebug(info->vm_info, info, "New guest CR4=%x and shadow CR4=%x\n", *(uint_t *)guest_cr4,*(uint_t*)shadow_cr4);
- shadow_pt = create_new_shadow_pt32();
+ } else if ((cpu_mode == LONG) || (cpu_mode == LONG_32_COMPAT)) {
+ struct cr4_64 * new_cr4 = (struct cr4_64 *)(dec_instr.src_operand.operand);
+ struct cr4_64 * cr4 = (struct cr4_64 *)&(info->ctrl_regs.cr4);
- shadow_cr3->pdt_base_addr = PD32_BASE_ADDR(shadow_pt);
- } else {
- PrintDebug("Reusing cached shadow Page table\n");
- }
-
- shadow_cr3->pwt = new_cr3->pwt;
- shadow_cr3->pcd = new_cr3->pcd;
-
- // What the hell...
- *guest_cr3 = *new_cr3;
-
- PrintDebug("New Shadow CR3=%x; New Guest CR3=%x\n",
- *(uint_t*)shadow_cr3, *(uint_t*)guest_cr3);
+ PrintDebug(info->vm_info, info, "Old CR4=%p\n", (void *)*(addr_t *)cr4);
+ PrintDebug(info->vm_info, info, "New CR4=%p\n", (void *)*(addr_t *)new_cr4);
+
+ if (new_cr4->pae == 0) {
+ // cannot turn off PAE in long mode GPF the guest
+ PrintError(info->vm_info, info, "Cannot disable PAE in long mode, should send GPF\n");
+ return -1;
+ }
+
+ *cr4 = *new_cr4;
+
+ } else {
+ PrintError(info->vm_info, info, "CR4 write not supported in CPU_MODE: %s\n", v3_cpu_mode_to_str(cpu_mode));
+ return -1;
+ }
+
+ if (info->shdw_pg_mode == SHADOW_PAGING) {
+ if (flush_tlb) {
+ PrintDebug(info->vm_info, info, "Handling PSE/PGE/PAE -> TLBFlush (doing flush now!)\n");
+ if (v3_activate_shadow_pt(info) == -1) {
+ PrintError(info->vm_info, info, "Failed to activate shadow page tables when emulating TLB flush in handling cr4 write\n");
+ return -1;
+ }
+ }
+ }
+
+ info->rip += dec_instr.instr_length;
+ return 0;
+}
- if (info->mem_mode == VIRTUAL_MEM) {
- // If we aren't in paged mode then we have to preserve the identity mapped CR3
- info->ctrl_regs.cr3 = *(addr_t*)shadow_cr3;
- }
+
+/*
+ The CR8 and APIC TPR interaction are kind of crazy.
+
+ CR8 mandates that the priority class is in bits 3:0
+
+ The interaction of CR8 and an actual APIC is somewhat implementation dependent, but
+ a basic current APIC has the priority class at 7:4 and the *subclass* at 3:0
+
+ The APIC TPR (both fields) can be written as the APIC register
+ A write to CR8 sets the priority class field, and should zero the subclass
+ A read from CR8 gets just the priority class field
+
+ In the apic_tpr storage location, we have:
+
+ zeros [class] [subclass]
+
+ Because of this, an APIC implementation should use apic_tpr to store its TPR
+ In fact, it *should* do this, otherwise its TPR may get out of sync with the architected TPR
+
+ On a CR8 read, we return just
+
+ zeros 0000 [class]
+
+ On a CR8 write, we set the register to
+
+ zeros [class] 0000
+
+*/
+
+int v3_handle_cr8_write(struct guest_info * info) {
+ int ret;
+ uchar_t instr[15];
+ struct x86_instr dec_instr;
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
}
- } else {
- PrintError("Unhandled opcode in handle_cr3_write\n");
- return -1;
- }
+
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
+ }
+
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
+
+ if (dec_instr.op_type == V3_OP_MOV2CR) {
+ PrintDebug(info->vm_info, info, "MOV2CR8 (cpu_mode=%s)\n", v3_cpu_mode_to_str(info->cpu_mode));
+
+ if ((info->cpu_mode == LONG) ||
+ (info->cpu_mode == LONG_32_COMPAT)) {
+ uint64_t *val = (uint64_t *)(dec_instr.src_operand.operand);
- info->rip += dec_instr.instr_length;
+ info->ctrl_regs.apic_tpr = (*val & 0xf) << 4;
- return 0;
+ V3_Print(info->vm_info, info, "Write of CR8 sets apic_tpr to 0x%llx\n",info->ctrl_regs.apic_tpr);
+
+ } else {
+ // probably should raise exception here
+ }
+ } else {
+ PrintError(info->vm_info, info, "Unhandled opcode in handle_cr8_write\n");
+ return -1;
+ }
+
+ info->rip += dec_instr.instr_length;
+
+ return 0;
}
-// first attempt = 156 lines
-// current = 36 lines
-int handle_cr3_read(struct guest_info * info) {
- char instr[15];
- int ret;
- struct x86_instr dec_instr;
-
- if (info->mem_mode == PHYSICAL_MEM) {
- ret = read_guest_pa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- } else {
- ret = read_guest_va_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
- }
-
- /* The IFetch will already have faulted in the necessary bytes for the full instruction
- if (ret != 15) {
- // I think we should inject a GPF into the guest
- PrintError("Could not read instruction (ret=%d)\n", ret);
- return -1;
- }
- */
-
- if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
- PrintError("Could not decode instruction\n");
- return -1;
- }
-
- if (opcode_cmp(V3_OPCODE_MOVCR2, (const uchar_t *)(dec_instr.opcode)) == 0) {
- PrintDebug("MOVCR32\n");
- struct cr3_32 * virt_cr3 = (struct cr3_32 *)(dec_instr.dst_operand.operand);
-
- PrintDebug("CR3 at 0x%x\n", &(info->ctrl_regs.cr3));
+int v3_handle_cr8_read(struct guest_info * info) {
+ uchar_t instr[15];
+ int ret;
+ struct x86_instr dec_instr;
+
+ if (info->mem_mode == PHYSICAL_MEM) {
+ ret = v3_read_gpa_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ } else {
+ ret = v3_read_gva_memory(info, get_addr_linear(info, info->rip, &(info->segments.cs)), 15, instr);
+ }
+
+ if (ret!=15) {
+ PrintError(info->vm_info, info, "Could not read instruction\n");
+ return -1;
+ }
+
+ if (v3_decode(info, (addr_t)instr, &dec_instr) == -1) {
+ PrintError(info->vm_info, info, "Could not decode instruction\n");
+ return -1;
+ }
+
+ if (dec_instr.op_type == V3_OP_MOVCR2) {
+ PrintDebug(info->vm_info, info, "MOVCR82 (mode=%s)\n", v3_cpu_mode_to_str(info->cpu_mode));
+
+ if ((info->cpu_mode == LONG) ||
+ (info->cpu_mode == LONG_32_COMPAT)) {
+ uint64_t *dst_reg = (uint64_t *)(dec_instr.dst_operand.operand);
- if (info->shdw_pg_mode == SHADOW_PAGING) {
- *virt_cr3 = *(struct cr3_32 *)&(info->shdw_pg_state.guest_cr3);
+ *dst_reg = (info->ctrl_regs.apic_tpr >> 4) & 0xf;
+
+ V3_Print(info->vm_info, info, "Read of CR8 (apic_tpr) returns 0x%llx\n",*dst_reg);
+
+ } else {
+ // probably should raise exception
+ }
+
} else {
- *virt_cr3 = *(struct cr3_32 *)&(info->ctrl_regs.cr3);
+ PrintError(info->vm_info, info, "Unhandled opcode in handle_cr8_read\n");
+ return -1;
}
- } else {
- PrintError("Unhandled opcode in handle_cr3_read\n");
- return -1;
- }
+
+ info->rip += dec_instr.instr_length;
+
+ return 0;
+}
- info->rip += dec_instr.instr_length;
- return 0;
+int v3_handle_efer_read(struct guest_info * core, uint_t msr, struct v3_msr * dst, void * priv_data) {
+ PrintDebug(core->vm_info, core, "EFER Read HI=%x LO=%x\n", core->shdw_pg_state.guest_efer.hi, core->shdw_pg_state.guest_efer.lo);
+
+ dst->value = core->shdw_pg_state.guest_efer.value;
+
+ return 0;
+}
+
+
+int v3_handle_efer_write(struct guest_info * core, uint_t msr, struct v3_msr src, void * priv_data) {
+ struct v3_msr * vm_efer = &(core->shdw_pg_state.guest_efer);
+ struct efer_64 * hw_efer = (struct efer_64 *)&(core->ctrl_regs.efer);
+ struct efer_64 old_hw_efer = *((struct efer_64 *)&core->ctrl_regs.efer);
+
+ PrintDebug(core->vm_info, core, "EFER Write HI=%x LO=%x\n", src.hi, src.lo);
+
+ // Set EFER value seen by guest if it reads EFER
+ vm_efer->value = src.value;
+
+ // Set EFER value seen by hardware while the guest is running
+ *(uint64_t *)hw_efer = src.value;
+
+ // We have gotten here either because we are using
+ // shadow paging, or we are using nested paging on SVM
+ // In the latter case, we don't need to do anything
+ // like the following
+ if (core->shdw_pg_mode == SHADOW_PAGING) {
+ // Catch unsupported features
+ if ((old_hw_efer.lme == 1) && (hw_efer->lme == 0)) {
+ PrintError(core->vm_info, core, "Disabling long mode once it has been enabled is not supported\n");
+ return -1;
+ }
+
+ // Set LME and LMA bits seen by hardware
+ if (old_hw_efer.lme == 0) {
+ // Long mode was not previously enabled, so the lme bit cannot
+ // be set yet. It will be set later when the guest sets CR0.PG
+ // to enable paging.
+ hw_efer->lme = 0;
+ } else {
+ // Long mode was previously enabled. Ensure LMA bit is set.
+ // VMX does not automatically set LMA, and this should not affect SVM.
+ hw_efer->lma = 1;
+ }
+ }
+
+
+ PrintDebug(core->vm_info, core, "RIP=%p\n", (void *)core->rip);
+ PrintDebug(core->vm_info, core, "New EFER value HW(hi=%p), VM(hi=%p)\n", (void *)*(uint64_t *)hw_efer, (void *)vm_efer->value);
+
+
+ return 0;
+}
+
+int v3_handle_vm_cr_read(struct guest_info * core, uint_t msr, struct v3_msr * dst, void * priv_data) {
+ /* tell the guest that the BIOS disabled SVM, that way it doesn't get
+ * confused by the fact that CPUID reports SVM as available but it still
+ * cannot be used
+ */
+ dst->value = SVM_VM_CR_MSR_lock | SVM_VM_CR_MSR_svmdis;
+ PrintDebug(core->vm_info, core, "VM_CR Read HI=%x LO=%x\n", dst->hi, dst->lo);
+ return 0;
+}
+
+int v3_handle_vm_cr_write(struct guest_info * core, uint_t msr, struct v3_msr src, void * priv_data) {
+ PrintDebug(core->vm_info, core, "VM_CR Write\n");
+ PrintDebug(core->vm_info, core, "VM_CR Write Values: HI=%x LO=%x\n", src.hi, src.lo);
+
+ /* writes to LOCK and SVMDIS are silently ignored (according to the spec),
+ * other writes indicate the guest wants to use some feature we haven't
+ * implemented
+ */
+ if (src.value & ~(SVM_VM_CR_MSR_lock | SVM_VM_CR_MSR_svmdis)) {
+ PrintDebug(core->vm_info, core, "VM_CR write sets unsupported bits: HI=%x LO=%x\n", src.hi, src.lo);
+ return -1;
+ }
+
+ return 0;
}
