[PATCH 3.13 105/162] KVM: x86: Handle errors when RIP is set during far jumps

From: Kamal Mostafa
Date: Thu Nov 06 2014 - 18:00:29 EST


3.13.11.11 -stable review patch. If anyone has any objections, please let me know.

------------------

From: Nadav Amit <namit@xxxxxxxxxxxxxxxxx>

commit d1442d85cc30ea75f7d399474ca738e0bc96f715 upstream.

Far jmp/call/ret may fault while loading a new RIP. Currently KVM does not
handle this case, and may result in failed vm-entry once the assignment is
done. The tricky part of doing so is that loading the new CS affects the
VMCS/VMCB state, so if we fail during loading the new RIP, we are left in
unconsistent state. Therefore, this patch saves on 64-bit the old CS
descriptor and restores it if loading RIP failed.

This fixes CVE-2014-3647.

Signed-off-by: Nadav Amit <namit@xxxxxxxxxxxxxxxxx>
Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>
[ kamal: backport to 3.13-stable: context; __long_segment_descriptor args ]
Signed-off-by: Kamal Mostafa <kamal@xxxxxxxxxxxxx>
---
arch/x86/kvm/emulate.c | 117 ++++++++++++++++++++++++++++++++++++-------------
1 file changed, 87 insertions(+), 30 deletions(-)

diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index 300c164..a440bea 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -1437,7 +1437,8 @@ static int write_segment_descriptor(struct x86_emulate_ctxt *ctxt,

/* Does not support long mode */
static int __load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
- u16 selector, int seg, u8 cpl)
+ u16 selector, int seg, u8 cpl,
+ struct desc_struct *desc)
{
struct desc_struct seg_desc, old_desc;
u8 dpl, rpl;
@@ -1563,6 +1564,8 @@ static int __load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
}
load:
ctxt->ops->set_segment(ctxt, selector, &seg_desc, 0, seg);
+ if (desc)
+ *desc = seg_desc;
return X86EMUL_CONTINUE;
exception:
emulate_exception(ctxt, err_vec, err_code, true);
@@ -1573,7 +1576,7 @@ static int load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
u16 selector, int seg)
{
u8 cpl = ctxt->ops->cpl(ctxt);
- return __load_segment_descriptor(ctxt, selector, seg, cpl);
+ return __load_segment_descriptor(ctxt, selector, seg, cpl, NULL);
}

static void write_register_operand(struct operand *op)
@@ -1970,17 +1973,31 @@ static int em_iret(struct x86_emulate_ctxt *ctxt)
static int em_jmp_far(struct x86_emulate_ctxt *ctxt)
{
int rc;
- unsigned short sel;
+ unsigned short sel, old_sel;
+ struct desc_struct old_desc, new_desc;
+ const struct x86_emulate_ops *ops = ctxt->ops;
+ u8 cpl = ctxt->ops->cpl(ctxt);
+
+ /* Assignment of RIP may only fail in 64-bit mode */
+ if (ctxt->mode == X86EMUL_MODE_PROT64)
+ ops->get_segment(ctxt, &old_sel, &old_desc, NULL,
+ VCPU_SREG_CS);

memcpy(&sel, ctxt->src.valptr + ctxt->op_bytes, 2);

- rc = load_segment_descriptor(ctxt, sel, VCPU_SREG_CS);
+ rc = __load_segment_descriptor(ctxt, sel, VCPU_SREG_CS, cpl,
+ &new_desc);
if (rc != X86EMUL_CONTINUE)
return rc;

- ctxt->_eip = 0;
- memcpy(&ctxt->_eip, ctxt->src.valptr, ctxt->op_bytes);
- return X86EMUL_CONTINUE;
+ rc = assign_eip_far(ctxt, ctxt->src.val, new_desc.l);
+ if (rc != X86EMUL_CONTINUE) {
+ WARN_ON(!ctxt->mode != X86EMUL_MODE_PROT64);
+ /* assigning eip failed; restore the old cs */
+ ops->set_segment(ctxt, old_sel, &old_desc, 0, VCPU_SREG_CS);
+ return rc;
+ }
+ return rc;
}

static int em_grp45(struct x86_emulate_ctxt *ctxt)
@@ -2044,21 +2061,34 @@ static int em_ret(struct x86_emulate_ctxt *ctxt)
static int em_ret_far(struct x86_emulate_ctxt *ctxt)
{
int rc;
- unsigned long cs;
+ unsigned long eip, cs;
+ u16 old_cs;
int cpl = ctxt->ops->cpl(ctxt);
+ struct desc_struct old_desc, new_desc;
+ const struct x86_emulate_ops *ops = ctxt->ops;
+
+ if (ctxt->mode == X86EMUL_MODE_PROT64)
+ ops->get_segment(ctxt, &old_cs, &old_desc, NULL,
+ VCPU_SREG_CS);

- rc = emulate_pop(ctxt, &ctxt->_eip, ctxt->op_bytes);
+ rc = emulate_pop(ctxt, &eip, ctxt->op_bytes);
if (rc != X86EMUL_CONTINUE)
return rc;
- if (ctxt->op_bytes == 4)
- ctxt->_eip = (u32)ctxt->_eip;
rc = emulate_pop(ctxt, &cs, ctxt->op_bytes);
if (rc != X86EMUL_CONTINUE)
return rc;
/* Outer-privilege level return is not implemented */
if (ctxt->mode >= X86EMUL_MODE_PROT16 && (cs & 3) > cpl)
return X86EMUL_UNHANDLEABLE;
- rc = load_segment_descriptor(ctxt, (u16)cs, VCPU_SREG_CS);
+ rc = __load_segment_descriptor(ctxt, (u16)cs, VCPU_SREG_CS, 0,
+ &new_desc);
+ if (rc != X86EMUL_CONTINUE)
+ return rc;
+ rc = assign_eip_far(ctxt, eip, new_desc.l);
+ if (rc != X86EMUL_CONTINUE) {
+ WARN_ON(!ctxt->mode != X86EMUL_MODE_PROT64);
+ ops->set_segment(ctxt, old_cs, &old_desc, 0, VCPU_SREG_CS);
+ }
return rc;
}

@@ -2482,19 +2512,24 @@ static int load_state_from_tss16(struct x86_emulate_ctxt *ctxt,
* Now load segment descriptors. If fault happens at this stage
* it is handled in a context of new task
*/
- ret = __load_segment_descriptor(ctxt, tss->ldt, VCPU_SREG_LDTR, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->ldt, VCPU_SREG_LDTR, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->es, VCPU_SREG_ES, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->es, VCPU_SREG_ES, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->cs, VCPU_SREG_CS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->cs, VCPU_SREG_CS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->ss, VCPU_SREG_SS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->ss, VCPU_SREG_SS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->ds, VCPU_SREG_DS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->ds, VCPU_SREG_DS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;

@@ -2620,25 +2655,32 @@ static int load_state_from_tss32(struct x86_emulate_ctxt *ctxt,
* Now load segment descriptors. If fault happenes at this stage
* it is handled in a context of new task
*/
- ret = __load_segment_descriptor(ctxt, tss->ldt_selector, VCPU_SREG_LDTR, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->ldt_selector, VCPU_SREG_LDTR,
+ cpl, NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->es, VCPU_SREG_ES, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->es, VCPU_SREG_ES, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->cs, VCPU_SREG_CS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->cs, VCPU_SREG_CS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->ss, VCPU_SREG_SS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->ss, VCPU_SREG_SS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->ds, VCPU_SREG_DS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->ds, VCPU_SREG_DS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->fs, VCPU_SREG_FS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->fs, VCPU_SREG_FS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;
- ret = __load_segment_descriptor(ctxt, tss->gs, VCPU_SREG_GS, cpl);
+ ret = __load_segment_descriptor(ctxt, tss->gs, VCPU_SREG_GS, cpl,
+ NULL);
if (ret != X86EMUL_CONTINUE)
return ret;

@@ -2918,24 +2960,39 @@ static int em_call_far(struct x86_emulate_ctxt *ctxt)
u16 sel, old_cs;
ulong old_eip;
int rc;
+ struct desc_struct old_desc, new_desc;
+ const struct x86_emulate_ops *ops = ctxt->ops;
+ int cpl = ctxt->ops->cpl(ctxt);

- old_cs = get_segment_selector(ctxt, VCPU_SREG_CS);
old_eip = ctxt->_eip;
+ ops->get_segment(ctxt, &old_cs, &old_desc, NULL, VCPU_SREG_CS);

memcpy(&sel, ctxt->src.valptr + ctxt->op_bytes, 2);
- if (load_segment_descriptor(ctxt, sel, VCPU_SREG_CS))
+ rc = __load_segment_descriptor(ctxt, sel, VCPU_SREG_CS, cpl,
+ &new_desc);
+ if (rc != X86EMUL_CONTINUE)
return X86EMUL_CONTINUE;

- ctxt->_eip = 0;
- memcpy(&ctxt->_eip, ctxt->src.valptr, ctxt->op_bytes);
+ rc = assign_eip_far(ctxt, ctxt->src.val, new_desc.l);
+ if (rc != X86EMUL_CONTINUE)
+ goto fail;

ctxt->src.val = old_cs;
rc = em_push(ctxt);
if (rc != X86EMUL_CONTINUE)
- return rc;
+ goto fail;

ctxt->src.val = old_eip;
- return em_push(ctxt);
+ rc = em_push(ctxt);
+ /* If we failed, we tainted the memory, but the very least we should
+ restore cs */
+ if (rc != X86EMUL_CONTINUE)
+ goto fail;
+ return rc;
+fail:
+ ops->set_segment(ctxt, old_cs, &old_desc, 0, VCPU_SREG_CS);
+ return rc;
+
}

static int em_ret_near_imm(struct x86_emulate_ctxt *ctxt)
--
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/