80  * If nested=1, nested virtualization is supported, i.e., guests may use
81  * VMX and be a hypervisor for its own guests. If nested=0, guests may not
84 static bool __read_mostly nested = 0;  variable
85 module_param(nested, bool, S_IRUGO);
152  * single nested guest (L2), hence the name vmcs12. Any VMX implementation has
335  * for correct emulation of VMX (i.e., nested VMX) on this vcpu.
378 	 * non-nested (L1) guest, it always points to vmcs01. For a nested
420 	/* Support for a guest hypervisor (nested VMX) */
421 	struct nested_vmx nested;  member
583 	return to_vmx(vcpu)->nested.current_vmcs12;  in get_vmcs12()
1186 	/* When we are running a nested L2 guest and L1 specified for it a  in update_exception_bitmap()
1546  * Return the cr0 value that a nested guest would read. This is a combination
1578 		 * only update vmcs12->guest_cr0 on nested exit).  in vmx_fpu_deactivate()
1671  * checks whether in a nested guest, we need to inject them to L1 or L2.
1806  * counter, even if a nested guest (L2) is currently running.
1814 		to_vmx(vcpu)->nested.vmcs01_tsc_offset :  in vmx_read_l1_tsc()
1842 		to_vmx(vcpu)->nested.vmcs01_tsc_offset = offset;  in vmx_write_tsc_offset()
1859 		to_vmx(vcpu)->nested.vmcs01_tsc_offset += adjustment;  in vmx_adjust_tsc_offset()
1876  * instructions and MSRs (i.e., nested VMX). Nested VMX is disabled for
1877  * all guests if the "nested" module option is off, and can also be disabled
1882 	return nested && guest_cpuid_has_vmx(vcpu);  in nested_vmx_allowed()
1887  * returned for the various VMX controls MSRs when nested VMX is enabled.
1889  * valid during nested entry from L1 to L2.
1991  * If we allow our guest to use VMX instructions (i.e., nested VMX), we should
2074 		/* Currently, no nested ept or nested vpid */  in vmx_get_vmx_msr()
2643 	if (nested)  in hardware_setup()
3040 		 * So basically the check on whether to allow nested VMX  in vmx_set_cr4()
3045 	} else if (to_vmx(vcpu)->nested.vmxon)  in vmx_set_cr4()
3936  * In nested virtualization, check if L1 asked to exit on external interrupts.
4084 		if (to_vmx(vcpu)->nested.nested_run_pending ||  in vmx_interrupt_allowed()
4337 	if (to_vmx(vcpu)->nested.vmxon &&  in handle_set_cr0()
4932  * So we keep, in vmx->nested.vmcs02_pool, a cache of size VMCS02_POOL_SIZE
4942 	list_for_each_entry(item, &vmx->nested.vmcs02_pool, list)  in nested_get_current_vmcs02()
4943 		if (item->vmptr == vmx->nested.current_vmptr) {  in nested_get_current_vmcs02()
4944 			list_move(&item->list, &vmx->nested.vmcs02_pool);  in nested_get_current_vmcs02()
4948 	if (vmx->nested.vmcs02_num >= max(VMCS02_POOL_SIZE, 1)) {  in nested_get_current_vmcs02()
4950 		item = list_entry(vmx->nested.vmcs02_pool.prev,  in nested_get_current_vmcs02()
4952 		item->vmptr = vmx->nested.current_vmptr;  in nested_get_current_vmcs02()
4953 		list_move(&item->list, &vmx->nested.vmcs02_pool);  in nested_get_current_vmcs02()
4968 	item->vmptr = vmx->nested.current_vmptr;  in nested_get_current_vmcs02()
4969 	list_add(&(item->list), &(vmx->nested.vmcs02_pool));  in nested_get_current_vmcs02()
4970 	vmx->nested.vmcs02_num++;  in nested_get_current_vmcs02()
4978 	list_for_each_entry(item, &vmx->nested.vmcs02_pool, list)  in nested_free_vmcs02()
4983 			vmx->nested.vmcs02_num--;  in nested_free_vmcs02()
4996 	list_for_each_entry_safe(item, n, &vmx->nested.vmcs02_pool, list) {  in nested_free_all_saved_vmcss()
5002 	vmx->nested.vmcs02_num = 0;  in nested_free_all_saved_vmcss()
5044 	INIT_LIST_HEAD(&(vmx->nested.vmcs02_pool));  in handle_vmon()
5045 	vmx->nested.vmcs02_num = 0;  in handle_vmon()
5047 	vmx->nested.vmxon = true;  in handle_vmon()
5063 	if (!vmx->nested.vmxon) {  in nested_vmx_check_permission()
5084  * Free whatever needs to be freed from vmx->nested when L1 goes down, or
5089 	if (!vmx->nested.vmxon)  in free_nested()
5091 	vmx->nested.vmxon = false;  in free_nested()
5092 	if (vmx->nested.current_vmptr != -1ull) {  in free_nested()
5093 		kunmap(vmx->nested.current_vmcs12_page);  in free_nested()
5094 		nested_release_page(vmx->nested.current_vmcs12_page);  in free_nested()
5095 		vmx->nested.current_vmptr = -1ull;  in free_nested()
5096 		vmx->nested.current_vmcs12 = NULL;  in free_nested()
5099 	if (vmx->nested.apic_access_page) {  in free_nested()
5100 		nested_release_page(vmx->nested.apic_access_page);  in free_nested()
5101 		vmx->nested.apic_access_page = 0;  in free_nested()
5193 	if (to_vmx(vcpu)->nested.current_vmptr == -1ull) {  in nested_vmx_failValid()
5237 	if (vmptr == vmx->nested.current_vmptr) {  in handle_vmclear()
5238 		kunmap(vmx->nested.current_vmcs12_page);  in handle_vmclear()
5239 		nested_release_page(vmx->nested.current_vmcs12_page);  in handle_vmclear()
5240 		vmx->nested.current_vmptr = -1ull;  in handle_vmclear()
5241 		vmx->nested.current_vmcs12 = NULL;  in handle_vmclear()
5249 		 * possible that a nested vmx bug, not a guest hypervisor bug,  in handle_vmclear()
5345 	if (vmx->nested.current_vmptr == -1ull) {  in nested_vmx_check_vmcs12()
5499 	if (vmx->nested.current_vmptr != vmptr) {  in handle_vmptrld()
5517 		if (vmx->nested.current_vmptr != -1ull) {  in handle_vmptrld()
5518 			kunmap(vmx->nested.current_vmcs12_page);  in handle_vmptrld()
5519 			nested_release_page(vmx->nested.current_vmcs12_page);  in handle_vmptrld()
5522 		vmx->nested.current_vmptr = vmptr;  in handle_vmptrld()
5523 		vmx->nested.current_vmcs12 = new_vmcs12;  in handle_vmptrld()
5524 		vmx->nested.current_vmcs12_page = page;  in handle_vmptrld()
5548 				 (void *)&to_vmx(vcpu)->nested.current_vmptr,  in handle_vmptrst()
5735 	if (vmx->nested.nested_run_pending)  in nested_vmx_exit_handled()
5853 	if (vmx->nested.nested_run_pending)  in vmx_handle_exit()
5858 		vmx->nested.nested_run_pending = 1;  in vmx_handle_exit()
5860 		vmx->nested.nested_run_pending = 0;  in vmx_handle_exit()
6103 	if (is_guest_mode(vcpu) && !vmx->nested.nested_run_pending) {  in vmx_vcpu_run()
6344 	vmx->nested.current_vmptr = -1ull;
6345 	vmx->nested.current_vmcs12 = NULL;
6441 	if (func == 1 && nested)
6446  * prepare_vmcs02 is called when the L1 guest hypervisor runs its nested
6562 			if (vmx->nested.apic_access_page) /* shouldn't happen */
6563 				nested_release_page(vmx->nested.apic_access_page);
6564 			vmx->nested.apic_access_page =
6572 			if (!vmx->nested.apic_access_page)
6577 				  page_to_phys(vmx->nested.apic_access_page));
6640 			vmx->nested.vmcs01_tsc_offset + vmcs12->tsc_offset);
6642 		vmcs_write64(TSC_OFFSET, vmx->nested.vmcs01_tsc_offset);
6686  * nested_vmx_run() handles a nested entry, i.e., a VMLAUNCH or VMRESUME on L1
6687  * for running an L2 nested guest.
6704 	 * The nested entry process starts with enforcing various prerequisites
6779 	 * the nested entry.
6788 	vmx->nested.vmcs01_tsc_offset = vmcs_read64(TSC_OFFSET);
6811  * On a nested exit from L2 to L1, vmcs12.guest_cr0 might not be up-to-date
6848  * prepare_vmcs12 is part of what we need to do when the nested L2 guest exits
6941  * A part of what we need to when the nested L2 guest exits and we want to
6944  * This function is to be called not only on normal nested exit, but also on
6945  * a nested entry failure, as explained in Intel's spec, 3B.23.7 ("VM-Entry
7022  * Emulate an exit from nested guest (L2) to L1, i.e., prepare to run L1
7044 		nested_free_vmcs02(vmx, vmx->nested.current_vmptr);
7049 	vmcs_write64(TSC_OFFSET, vmx->nested.vmcs01_tsc_offset);
7055 	if (vmx->nested.apic_access_page) {
7056 		nested_release_page(vmx->nested.apic_access_page);
7057 		vmx->nested.apic_access_page = 0;