WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-devel

[Xen-devel] RFC: Nested VMX patch series 03: vmxon_off

To: "Dong, Eddie" <eddie.dong@xxxxxxxxx>, Tim Deegan <Tim.Deegan@xxxxxxxxxx>, Keir Fraser <keir@xxxxxxx>
Subject: [Xen-devel] RFC: Nested VMX patch series 03: vmxon_off
From: "Dong, Eddie" <eddie.dong@xxxxxxxxx>
Date: Wed, 1 Jun 2011 11:52:34 +0800
Accept-language: en-US
Acceptlanguage: en-US
Cc: "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>, "Dong, Eddie" <eddie.dong@xxxxxxxxx>, "He, Qing" <qing.he@xxxxxxxxx>
Delivery-date: Tue, 31 May 2011 20:55:11 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
References: <osstest-7468-mainreport@xxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: AcwgAhjDwUdZ/2BOTBqtK+IA8ti/WgAC9edgAAAdViAAAA3MIAAAIerw
Thread-topic: [Xen-devel] RFC: Nested VMX patch series 03: vmxon_off
VMXON_OFF processing

Thx, Eddie

        Signed-off-by: Qing He <qing.he@xxxxxxxxx>
        Signed-off-by: Eddie Dong <eddie.dong@xxxxxxxxx>

diff -r 7a9c7fbdabc0 xen/arch/x86/hvm/vmx/Makefile
--- a/xen/arch/x86/hvm/vmx/Makefile     Wed Jun 01 08:48:22 2011 +0800
+++ b/xen/arch/x86/hvm/vmx/Makefile     Wed Jun 01 09:39:53 2011 +0800
@@ -5,3 +5,4 @@
 obj-y += vmx.o
 obj-y += vpmu_core2.o
 obj-y += vvmx.o
+obj-y += vvmx.o
diff -r 7a9c7fbdabc0 xen/arch/x86/hvm/vmx/vmx.c
--- a/xen/arch/x86/hvm/vmx/vmx.c        Wed Jun 01 08:48:22 2011 +0800
+++ b/xen/arch/x86/hvm/vmx/vmx.c        Wed Jun 01 09:39:53 2011 +0800
@@ -2435,6 +2435,16 @@
         break;
     }
 
+    case EXIT_REASON_VMXOFF:
+        if ( nvmx_handle_vmxoff(regs) == X86EMUL_OKAY )
+            update_guest_eip();
+        break;
+
+    case EXIT_REASON_VMXON:
+        if ( nvmx_handle_vmxon(regs) == X86EMUL_OKAY )
+            update_guest_eip();
+        break;
+
     case EXIT_REASON_MWAIT_INSTRUCTION:
     case EXIT_REASON_MONITOR_INSTRUCTION:
     case EXIT_REASON_VMCLEAR:
@@ -2444,8 +2454,6 @@
     case EXIT_REASON_VMREAD:
     case EXIT_REASON_VMRESUME:
     case EXIT_REASON_VMWRITE:
-    case EXIT_REASON_VMXOFF:
-    case EXIT_REASON_VMXON:
     case EXIT_REASON_GETSEC:
     case EXIT_REASON_INVEPT:
     case EXIT_REASON_INVVPID:
diff -r 7a9c7fbdabc0 xen/arch/x86/hvm/vmx/vvmx.c
--- a/xen/arch/x86/hvm/vmx/vvmx.c       Wed Jun 01 08:48:22 2011 +0800
+++ b/xen/arch/x86/hvm/vmx/vvmx.c       Wed Jun 01 09:39:53 2011 +0800
@@ -91,3 +91,281 @@
     return 0;
 }
 
+/*
+ * Encode of VMX instructions base on Table 24-11 & 24-12 of SDM 3B
+ */
+
+enum vmx_regs_enc {
+    VMX_REG_RAX,
+    VMX_REG_RCX,
+    VMX_REG_RDX,
+    VMX_REG_RBX,
+    VMX_REG_RSP,
+    VMX_REG_RBP,
+    VMX_REG_RSI,
+    VMX_REG_RDI,
+#ifdef CONFIG_X86_64
+    VMX_REG_R8,
+    VMX_REG_R9,
+    VMX_REG_R10,
+    VMX_REG_R11,
+    VMX_REG_R12,
+    VMX_REG_R13,
+    VMX_REG_R14,
+    VMX_REG_R15,
+#endif
+};
+
+enum vmx_sregs_enc {
+    VMX_SREG_ES,
+    VMX_SREG_CS,
+    VMX_SREG_SS,
+    VMX_SREG_DS,
+    VMX_SREG_FS,
+    VMX_SREG_GS,
+};
+
+union vmx_inst_info {
+    struct {
+        unsigned int scaling           :2; /* bit 0-1 */
+        unsigned int __rsvd0           :1; /* bit 2 */
+        unsigned int reg1              :4; /* bit 3-6 */
+        unsigned int addr_size         :3; /* bit 7-9 */
+        unsigned int memreg            :1; /* bit 10 */
+        unsigned int __rsvd1           :4; /* bit 11-14 */
+        unsigned int segment           :3; /* bit 15-17 */
+        unsigned int index_reg         :4; /* bit 18-21 */
+        unsigned int index_reg_invalid :1; /* bit 22 */
+        unsigned int base_reg          :4; /* bit 23-26 */
+        unsigned int base_reg_invalid  :1; /* bit 27 */
+        unsigned int reg2              :4; /* bit 28-31 */
+    } fields;
+    u32 word;
+};
+
+enum x86_segment sreg_to_index[] = {
+    [VMX_SREG_ES] = x86_seg_es,
+    [VMX_SREG_CS] = x86_seg_cs,
+    [VMX_SREG_SS] = x86_seg_ss,
+    [VMX_SREG_DS] = x86_seg_ds,
+    [VMX_SREG_FS] = x86_seg_fs,
+    [VMX_SREG_GS] = x86_seg_gs,
+};
+
+struct vmx_inst_decoded {
+#define VMX_INST_MEMREG_TYPE_MEMORY 0
+#define VMX_INST_MEMREG_TYPE_REG    1
+    int type;
+    union {
+        struct {
+            unsigned long mem;
+            unsigned int  len;
+        };
+        enum vmx_regs_enc reg1;
+    };
+
+    enum vmx_regs_enc reg2;
+};
+
+enum vmx_ops_result {
+    VMSUCCEED,
+    VMFAIL_VALID,
+    VMFAIL_INVALID,
+};
+
+#define CASE_GET_REG(REG, reg)      \
+    case VMX_REG_ ## REG: value = regs->reg; break
+
+static unsigned long reg_read(struct cpu_user_regs *regs,
+                              enum vmx_regs_enc index)
+{
+    unsigned long value = 0;
+
+    switch ( index ) {
+    CASE_GET_REG(RAX, eax);
+    CASE_GET_REG(RCX, ecx);
+    CASE_GET_REG(RDX, edx);
+    CASE_GET_REG(RBX, ebx);
+    CASE_GET_REG(RBP, ebp);
+    CASE_GET_REG(RSI, esi);
+    CASE_GET_REG(RDI, edi);
+    CASE_GET_REG(RSP, esp);
+#ifdef CONFIG_X86_64
+    CASE_GET_REG(R8, r8);
+    CASE_GET_REG(R9, r9);
+    CASE_GET_REG(R10, r10);
+    CASE_GET_REG(R11, r11);
+    CASE_GET_REG(R12, r12);
+    CASE_GET_REG(R13, r13);
+    CASE_GET_REG(R14, r14);
+    CASE_GET_REG(R15, r15);
+#endif
+    default:
+        break;
+    }
+
+    return value;
+}
+
+static int vmx_inst_check_privilege(struct cpu_user_regs *regs, int 
vmxop_check)
+{
+    struct vcpu *v = current;
+    struct segment_register cs;
+
+    hvm_get_segment_register(v, x86_seg_cs, &cs);
+
+    if ( vmxop_check )
+    {
+        if ( !(v->arch.hvm_vcpu.guest_cr[0] & X86_CR0_PE) ||
+             !(v->arch.hvm_vcpu.guest_cr[4] & X86_CR4_VMXE) )
+            goto invalid_op;
+    }
+    else if ( !vcpu_2_nvmx(v).vmxon_region_pa )
+        goto invalid_op;
+
+    if ( (regs->eflags & X86_EFLAGS_VM) ||
+         (hvm_long_mode_enabled(v) && cs.attr.fields.l == 0) )
+        goto invalid_op;
+    /* TODO: check vmx operation mode */
+
+    if ( (cs.sel & 3) > 0 )
+        goto gp_fault;
+
+    return X86EMUL_OKAY;
+
+invalid_op:
+    gdprintk(XENLOG_ERR, "vmx_inst_check_privilege: invalid_op\n");
+    hvm_inject_exception(TRAP_invalid_op, 0, 0);
+    return X86EMUL_EXCEPTION;
+
+gp_fault:
+    gdprintk(XENLOG_ERR, "vmx_inst_check_privilege: gp_fault\n");
+    hvm_inject_exception(TRAP_gp_fault, 0, 0);
+    return X86EMUL_EXCEPTION;
+}
+
+static int decode_vmx_inst(struct cpu_user_regs *regs,
+                           struct vmx_inst_decoded *decode,
+                           unsigned long *poperandS, int vmxon_check)
+{
+    struct vcpu *v = current;
+    union vmx_inst_info info;
+    struct segment_register seg;
+    unsigned long base, index, seg_base, disp, offset;
+    int scale, size;
+
+    if ( vmx_inst_check_privilege(regs, vmxon_check) != X86EMUL_OKAY )
+        return X86EMUL_EXCEPTION;
+
+    info.word = __vmread(VMX_INSTRUCTION_INFO);
+
+    if ( info.fields.memreg ) {
+        decode->type = VMX_INST_MEMREG_TYPE_REG;
+        decode->reg1 = info.fields.reg1;
+        if ( poperandS != NULL )
+            *poperandS = reg_read(regs, decode->reg1);
+    }
+    else
+    {
+        decode->type = VMX_INST_MEMREG_TYPE_MEMORY;
+        hvm_get_segment_register(v, sreg_to_index[info.fields.segment], &seg);
+        /* TODO: segment type check */
+        seg_base = seg.base;
+
+        base = info.fields.base_reg_invalid ? 0 :
+            reg_read(regs, info.fields.base_reg);
+
+        index = info.fields.index_reg_invalid ? 0 :
+            reg_read(regs, info.fields.index_reg);
+
+        scale = 1 << info.fields.scaling;
+
+        disp = __vmread(EXIT_QUALIFICATION);
+
+        size = 1 << (info.fields.addr_size + 1);
+
+        offset = base + index * scale + disp;
+        if ( (offset > seg.limit || offset + size > seg.limit) &&
+            (!hvm_long_mode_enabled(v) || info.fields.segment == VMX_SREG_GS) )
+            goto gp_fault;
+
+        if ( poperandS != NULL &&
+             hvm_copy_from_guest_virt(poperandS, seg_base + offset, size, 0)
+                  != HVMCOPY_okay )
+            return X86EMUL_EXCEPTION;
+        decode->mem = seg_base + offset;
+        decode->len = size;
+    }
+
+    decode->reg2 = info.fields.reg2;
+
+    return X86EMUL_OKAY;
+
+gp_fault:
+    hvm_inject_exception(TRAP_gp_fault, 0, 0);
+    return X86EMUL_EXCEPTION;
+}
+
+static void vmreturn(struct cpu_user_regs *regs, enum vmx_ops_result ops_res)
+{
+    unsigned long eflags = regs->eflags;
+    unsigned long mask = X86_EFLAGS_CF | X86_EFLAGS_PF | X86_EFLAGS_AF |
+                         X86_EFLAGS_ZF | X86_EFLAGS_SF | X86_EFLAGS_OF;
+
+    eflags &= ~mask;
+
+    switch ( ops_res ) {
+    case VMSUCCEED:
+        break;
+    case VMFAIL_VALID:
+        /* TODO: error number, useful for guest VMM debugging */
+        eflags |= X86_EFLAGS_ZF;
+        break;
+    case VMFAIL_INVALID:
+    default:
+        eflags |= X86_EFLAGS_CF;
+        break;
+    }
+
+    regs->eflags = eflags;
+}
+
+/*
+ * VMX instructions handling
+ */
+
+int nvmx_handle_vmxon(struct cpu_user_regs *regs)
+{
+    struct vcpu *v=current;
+    struct nestedvmx *nvmx = &vcpu_2_nvmx(v);
+    struct vmx_inst_decoded decode;
+    unsigned long gpa = 0;
+    int rc;
+
+    rc = decode_vmx_inst(regs, &decode, &gpa, 1);
+    if ( rc != X86EMUL_OKAY )
+        return rc;
+
+    nvmx->vmxon_region_pa = gpa;
+    vmreturn(regs, VMSUCCEED);
+
+    return X86EMUL_OKAY;
+
+}
+
+int nvmx_handle_vmxoff(struct cpu_user_regs *regs)
+{
+    struct vcpu *v=current;
+    struct nestedvmx *nvmx = &vcpu_2_nvmx(v);
+    int rc;
+
+    rc = vmx_inst_check_privilege(regs, 0);
+    if ( rc != X86EMUL_OKAY )
+        return rc;
+
+    nvmx->vmxon_region_pa = 0;
+
+    vmreturn(regs, VMSUCCEED);
+    return X86EMUL_OKAY;
+}
+
diff -r 7a9c7fbdabc0 xen/include/asm-x86/hvm/vmx/vvmx.h
--- a/xen/include/asm-x86/hvm/vmx/vvmx.h        Wed Jun 01 08:48:22 2011 +0800
+++ b/xen/include/asm-x86/hvm/vmx/vvmx.h        Wed Jun 01 09:39:53 2011 +0800
@@ -46,5 +46,7 @@
 uint32_t nvmx_vcpu_asid(struct vcpu *v);
 enum hvm_intblk nvmx_intr_blocked(struct vcpu *v);
 
+int nvmx_handle_vmxon(struct cpu_user_regs *regs);
+int nvmx_handle_vmxoff(struct cpu_user_regs *regs);
 #endif /* __ASM_X86_HVM_VVMX_H__ */


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel