From: Mihai Donțu <mdontu@xxxxxxxxxxxxxxx>
This command is used to get kvm_regs and kvm_sregs structures,
plus a list of struct kvm_msrs from a specific vCPU.
While the kvm_regs and kvm_sregs structures are included with every
event, this command allows reading any MSR and can be used as a quick
way to read the state of any vCPU.
Signed-off-by: Mihai Donțu <mdontu@xxxxxxxxxxxxxxx>
Co-developed-by: Adalbert Lazăr <alazar@xxxxxxxxxxxxxxx>
Signed-off-by: Adalbert Lazăr <alazar@xxxxxxxxxxxxxxx>
---
Documentation/virt/kvm/kvmi.rst | 44 +++++++++
arch/x86/include/uapi/asm/kvmi.h | 15 +++
arch/x86/kvm/kvmi.c | 93 +++++++++++++++++++
include/uapi/linux/kvmi.h | 1 +
.../testing/selftests/kvm/x86_64/kvmi_test.c | 75 +++++++++++++++
virt/kvm/introspection/kvmi_int.h | 7 ++
virt/kvm/introspection/kvmi_msg.c | 20 ++++
7 files changed, 255 insertions(+)
diff --git a/Documentation/virt/kvm/kvmi.rst b/Documentation/virt/kvm/kvmi.rst
index 4393ce89b2fa..f9095e1a9417 100644
--- a/Documentation/virt/kvm/kvmi.rst
+++ b/Documentation/virt/kvm/kvmi.rst
@@ -568,6 +568,50 @@ the *KVMI_VM_CONTROL_EVENTS* command.
* -KVM_EPERM - the access is disallowed (use *KVMI_VM_CHECK_EVENT* first)
* -KVM_EAGAIN - the selected vCPU can't be introspected yet
+11. KVMI_VCPU_GET_REGISTERS
+---------------------------
+
+:Architectures: x86
+:Versions: >= 1
+:Parameters:
+
+::
+
+ struct kvmi_vcpu_hdr;
+ struct kvmi_vcpu_get_registers {
+ __u16 nmsrs;
+ __u16 padding1;
+ __u32 padding2;
+ __u32 msrs_idx[0];
+ };
+
+:Returns:
+
+::
+
+ struct kvmi_error_code;
+ struct kvmi_vcpu_get_registers_reply {
+ __u32 mode;
+ __u32 padding;
+ struct kvm_regs regs;
+ struct kvm_sregs sregs;
+ struct kvm_msrs msrs;
+ };
+
+For the given vCPU and the ``nmsrs`` sized array of MSRs registers,
+returns the current vCPU mode (in bytes: 2, 4 or 8), the general purpose
+registers, the special registers and the requested set of MSRs.
+
+:Errors:
+
+* -KVM_EINVAL - the selected vCPU is invalid
+* -KVM_EINVAL - one of the indicated MSRs is invalid
+* -KVM_EINVAL - the padding is not zero
+* -KVM_EINVAL - the reply size is larger than KVMI_MSG_SIZE
+ (too many MSRs)
+* -KVM_EAGAIN - the selected vCPU can't be introspected yet
+* -KVM_ENOMEM - there is not enough memory to allocate the reply
+
Events
======
diff --git a/arch/x86/include/uapi/asm/kvmi.h b/arch/x86/include/uapi/asm/kvmi.h
index 89adf84cefe4..f14674c3c109 100644
--- a/arch/x86/include/uapi/asm/kvmi.h
+++ b/arch/x86/include/uapi/asm/kvmi.h
@@ -30,4 +30,19 @@ struct kvmi_vcpu_get_info_reply {
__u64 tsc_speed;
};
+struct kvmi_vcpu_get_registers {
+ __u16 nmsrs;
+ __u16 padding1;
+ __u32 padding2;
+ __u32 msrs_idx[0];
+};
+
+struct kvmi_vcpu_get_registers_reply {
+ __u32 mode;
+ __u32 padding;
+ struct kvm_regs regs;
+ struct kvm_sregs sregs;
+ struct kvm_msrs msrs;
+};
+
#endif /* _UAPI_ASM_X86_KVMI_H */
diff --git a/arch/x86/kvm/kvmi.c b/arch/x86/kvm/kvmi.c
index ce7e2d5f2ab4..4fd7a3c17ef5 100644
--- a/arch/x86/kvm/kvmi.c
+++ b/arch/x86/kvm/kvmi.c
@@ -98,3 +98,96 @@ int kvmi_arch_cmd_vcpu_get_info(struct kvm_vcpu *vcpu,
return 0;
}
+
+int kvmi_arch_check_get_registers_req(const struct kvmi_msg_hdr *msg,
+ const struct kvmi_vcpu_get_registers *req)
+{
+ size_t req_size;
+
+ if (check_add_overflow(sizeof(struct kvmi_vcpu_hdr),
+ struct_size(req, msrs_idx, req->nmsrs),
+ &req_size))
+ return -1;
+
+ if (msg->size < req_size)
+ return -1;
+
+ return 0;
+}
+
+static int kvmi_get_registers(struct kvm_vcpu *vcpu, u32 *mode,
+ struct kvm_regs *regs,
+ struct kvm_sregs *sregs,
+ struct kvm_msrs *msrs)
+{
+ struct kvm_msr_entry *msr = msrs->entries;
+ struct kvm_msr_entry *end = msrs->entries + msrs->nmsrs;
+ struct msr_data m = {.host_initiated = true};
+ int err = 0;
+
+ kvm_arch_vcpu_get_regs(vcpu, regs);
+ kvm_arch_vcpu_get_sregs(vcpu, sregs);
+ *mode = kvmi_vcpu_mode(vcpu, sregs);
+
+ for (; msr < end && !err; msr++) {
+ m.index = msr->index;
+
+ err = kvm_x86_ops.get_msr(vcpu, &m);
+
+ if (!err)
+ msr->data = m.data;
+ }
+
+ return err ? -KVM_EINVAL : 0;
+}
+
+static bool valid_reply_size(size_t rpl_size)
+{
+ size_t msg_size;
+
+ if (check_add_overflow(sizeof(struct kvmi_error_code),
+ rpl_size, &msg_size))
+ return false;
+
+ if (msg_size > KVMI_MSG_SIZE)
+ return false;
+
+ return true;
+}
+
+int kvmi_arch_cmd_vcpu_get_registers(struct kvm_vcpu *vcpu,
+ const struct kvmi_msg_hdr *msg,
+ const struct kvmi_vcpu_get_registers *req,
+ struct kvmi_vcpu_get_registers_reply **dest,
+ size_t *dest_size)
+{
+ struct kvmi_vcpu_get_registers_reply *rpl;
+ size_t rpl_size;
+ int err;
+ u16 k;
+
+ if (req->padding1 || req->padding2)
+ return -KVM_EINVAL;
+
+ rpl_size = struct_size(rpl, msrs.entries, req->nmsrs);
+
+ if (!valid_reply_size(rpl_size))
+ return -KVM_EINVAL;
+
+ rpl = kvmi_msg_alloc();
+ if (!rpl)
+ return -KVM_ENOMEM;
+
+ rpl->msrs.nmsrs = req->nmsrs;
+
+ for (k = 0; k < req->nmsrs; k++)
+ rpl->msrs.entries[k].index = req->msrs_idx[k];
+
+ err = kvmi_get_registers(vcpu, &rpl->mode, &rpl->regs,
+ &rpl->sregs, &rpl->msrs);
+
+ *dest = rpl;
+ *dest_size = rpl_size;
+
+ return err;
+}
diff --git a/include/uapi/linux/kvmi.h b/include/uapi/linux/kvmi.h
index 9ebf17fa9564..39ff54b4b661 100644
--- a/include/uapi/linux/kvmi.h
+++ b/include/uapi/linux/kvmi.h
@@ -28,6 +28,7 @@ enum {
KVMI_VCPU_GET_INFO = 8,
KVMI_VCPU_PAUSE = 9,
KVMI_VCPU_CONTROL_EVENTS = 10,
+ KVMI_VCPU_GET_REGISTERS = 11,
KVMI_NUM_MESSAGES
};
diff --git a/tools/testing/selftests/kvm/x86_64/kvmi_test.c b/tools/testing/selftests/kvm/x86_64/kvmi_test.c
index da6a06fa0baa..73aafc5d959a 100644
--- a/tools/testing/selftests/kvm/x86_64/kvmi_test.c
+++ b/tools/testing/selftests/kvm/x86_64/kvmi_test.c
@@ -887,6 +887,80 @@ static void test_cmd_vcpu_control_events(struct kvm_vm *vm)
}
+static void cmd_vcpu_get_registers(struct kvm_vm *vm, struct kvm_regs *regs)
+{
+ struct {
+ struct kvmi_msg_hdr hdr;
+ struct kvmi_vcpu_hdr vcpu_hdr;
+ struct kvmi_vcpu_get_registers cmd;
+ } req = {};
+ struct kvmi_vcpu_get_registers_reply rpl;
+
+ test_vcpu0_command(vm, KVMI_VCPU_GET_REGISTERS, &req.hdr, sizeof(req),
+ &rpl, sizeof(rpl));
+
+ memcpy(regs, &rpl.regs, sizeof(*regs));
+}
+
+static void test_invalid_cmd_vcpu_get_registers(struct kvm_vm *vm,
+ struct kvmi_msg_hdr *req,
+ size_t req_size, void *rpl,
+ size_t rpl_size)
+{
+ int r;
+
+ r = do_vcpu0_command(vm, KVMI_VCPU_GET_REGISTERS, req, req_size,
+ rpl, rpl_size);
+ TEST_ASSERT(r == -KVM_EINVAL,
+ "KVMI_VCPU_GET_REGISTERS didn't failed with -KVM_EINVAL, error %d (%s)\n",
+ -r, kvm_strerror(-r));
+}
+
+static void test_invalid_vcpu_get_registers(struct kvm_vm *vm)
+{
+ struct {
+ struct kvmi_msg_hdr hdr;
+ struct kvmi_vcpu_hdr vcpu_hdr;
+ struct kvmi_vcpu_get_registers cmd;
+ __u32 msrs_idx[1];
+ } req = {};
+ struct {
+ struct kvmi_msg_hdr hdr;
+ struct kvmi_vcpu_hdr vcpu_hdr;
+ struct kvmi_vcpu_get_registers cmd;
+ } *req_big;
+ struct kvmi_vcpu_get_registers_reply rpl;
+
+ req.cmd.padding1 = 1;
+ req.cmd.padding2 = 1;
+ test_invalid_cmd_vcpu_get_registers(vm, &req.hdr, sizeof(req),
+ &rpl, sizeof(rpl));
+
+ req.cmd.padding1 = 0;
+ req.cmd.padding2 = 0;
+ req.cmd.nmsrs = 1;
+ req.cmd.msrs_idx[0] = 0xffffffff;
+ test_invalid_cmd_vcpu_get_registers(vm, &req.hdr, sizeof(req),
+ &rpl, sizeof(rpl));
+
+ req_big = calloc(1, KVMI_MSG_SIZE);
+ req_big->cmd.nmsrs = (KVMI_MSG_SIZE - sizeof(*req_big)) / sizeof(__u32);
+ test_invalid_cmd_vcpu_get_registers(vm, &req.hdr, sizeof(req),
+ &rpl, sizeof(rpl));
+ free(req_big);
+}
+
+static void test_cmd_vcpu_get_registers(struct kvm_vm *vm)
+{
+ struct kvm_regs regs = {};
+
+ cmd_vcpu_get_registers(vm, ®s);
+
+ pr_info("get_registers rip 0x%llx\n", regs.rip);
+
+ test_invalid_vcpu_get_registers(vm);
+}
+
static void test_introspection(struct kvm_vm *vm)
{
srandom(time(0));
@@ -904,6 +978,7 @@ static void test_introspection(struct kvm_vm *vm)
test_cmd_vcpu_get_info(vm);
test_pause(vm);
test_cmd_vcpu_control_events(vm);
+ test_cmd_vcpu_get_registers(vm);
unhook_introspection(vm);
}
diff --git a/virt/kvm/introspection/kvmi_int.h b/virt/kvm/introspection/kvmi_int.h
index 57a62ebadd94..740eea3a9531 100644
--- a/virt/kvm/introspection/kvmi_int.h
+++ b/virt/kvm/introspection/kvmi_int.h
@@ -57,5 +57,12 @@ int kvmi_cmd_vcpu_pause(struct kvm_vcpu *vcpu, bool wait);
int kvmi_arch_cmd_vcpu_get_info(struct kvm_vcpu *vcpu,
struct kvmi_vcpu_get_info_reply *rpl);
void kvmi_arch_setup_event(struct kvm_vcpu *vcpu, struct kvmi_event *ev);
+int kvmi_arch_check_get_registers_req(const struct kvmi_msg_hdr *msg,
+ const struct kvmi_vcpu_get_registers *req);
+int kvmi_arch_cmd_vcpu_get_registers(struct kvm_vcpu *vcpu,
+ const struct kvmi_msg_hdr *msg,
+ const struct kvmi_vcpu_get_registers *req,
+ struct kvmi_vcpu_get_registers_reply **dest,
+ size_t *dest_size);
#endif
diff --git a/virt/kvm/introspection/kvmi_msg.c b/virt/kvm/introspection/kvmi_msg.c
index 20ef4a44d3a2..6c7a600dd477 100644
--- a/virt/kvm/introspection/kvmi_msg.c
+++ b/virt/kvm/introspection/kvmi_msg.c
@@ -418,6 +418,25 @@ static int handle_vcpu_control_events(const struct kvmi_vcpu_msg_job *job,
return kvmi_msg_vcpu_reply(job, msg, ec, NULL, 0);
}
+static int handle_vcpu_get_registers(const struct kvmi_vcpu_msg_job *job,
+ const struct kvmi_msg_hdr *msg,
+ const void *req)
+{
+ struct kvmi_vcpu_get_registers_reply *rpl = NULL;
+ size_t rpl_size = 0;
+ int err, ec;
+
+ if (kvmi_arch_check_get_registers_req(msg, req))
+ return -EINVAL;
+
+ ec = kvmi_arch_cmd_vcpu_get_registers(job->vcpu, msg, req,
+ &rpl, &rpl_size);
+
+ err = kvmi_msg_vcpu_reply(job, msg, ec, rpl, rpl_size);
+ kvmi_msg_free(rpl);
+ return err;
+}
+
/*
* These functions are executed from the vCPU thread. The receiving thread
* passes the messages using a newly allocated 'struct kvmi_vcpu_msg_job'
@@ -429,6 +448,7 @@ static int(*const msg_vcpu[])(const struct kvmi_vcpu_msg_job *,
[KVMI_EVENT] = handle_vcpu_event_reply,
[KVMI_VCPU_CONTROL_EVENTS] = handle_vcpu_control_events,
[KVMI_VCPU_GET_INFO] = handle_vcpu_get_info,
+ [KVMI_VCPU_GET_REGISTERS] = handle_vcpu_get_registers,
};
static bool is_vcpu_command(u16 id)
