[PATCH 0/5] KVM: selftests: Introduce vcpu_arch_put_guest()

[PATCH 0/5] KVM: selftests: Introduce vcpu_arch_put_guest()

[Sean Christopherson]

The end goal of this series is to add a regression test for commit
910c57dfa4d1 ("KVM: x86: Mark target gfn of emulated atomic instruction
as dirty"), *without* polluting the common dirty_log_test.c code with
gory x86 details.

The regression test requires forcing KVM to emulate a guest atomic RMW
access, which is done via a magic instruction prefix/opcode that is
guarded by an off-by-default KVM module param.

To allow x86 to (a) detect the param, (b) shove in the unique instruction,
and (c) do all of that conditionally so that selftests doesn't test _only_
the forced emulation path, this series provides a pseudo-RNG instance for
all tests, and a new arch hook for doing "interesting" guest writes
(vcpu_arch_put_guest()).

Tested on x86 and ARM, compile tested on s390 and RISC-V.

Sean Christopherson (5):
  KVM: selftests: Provide a global pseudo-RNG instance for all tests
  KVM: selftests: Provide an API for getting a random bool from an RNG
  KVM: selftests: Add global snapshot of
    kvm_is_forced_emulation_enabled()
  KVM: selftests: Add vcpu_arch_put_guest() to do writes from guest code
  KVM: selftests: Randomly force emulation on x86 writes from guest code

 .../selftests/kvm/dirty_log_perf_test.c       |  9 ++++----
 tools/testing/selftests/kvm/dirty_log_test.c  | 22 ++++--------------
 .../selftests/kvm/include/kvm_util_base.h     |  3 +++
 .../testing/selftests/kvm/include/memstress.h |  1 -
 .../testing/selftests/kvm/include/test_util.h | 19 +++++++++++++++
 .../kvm/include/x86_64/kvm_util_arch.h        | 23 +++++++++++++++++++
 tools/testing/selftests/kvm/lib/kvm_util.c    |  9 ++++++++
 tools/testing/selftests/kvm/lib/memstress.c   | 10 ++------
 .../selftests/kvm/lib/x86_64/processor.c      |  3 +++
 .../selftests/kvm/x86_64/pmu_counters_test.c  |  3 ---
 .../kvm/x86_64/userspace_msr_exit_test.c      | 10 ++------
 .../selftests/kvm/x86_64/xen_shinfo_test.c    |  5 ++--
 12 files changed, 72 insertions(+), 45 deletions(-)


base-commit: e9a2bba476c8332ed547fce485c158d03b0b9659
-- 
2.44.0.291.gc1ea87d7ee-goog

[PATCH 1/5] KVM: selftests: Provide a global pseudo-RNG instance for all tests

[Sean Christopherson]

Add a global guest_random_state instance, i.e. a pseudo-RNG, so that an
RNG is available for *all* tests.  This will allow randomizing behavior
in core library code, e.g. x86 will utilize the pRNG to conditionally
force emulation of writes from within common guest code.

To allow for deterministic runs, and to be compatible with existing tests,
allow tests to override the seed used to initialize the pRNG.

Note, the seed *must* be overwritten before a VM is created in order for
the seed to take effect, though it's perfectly fine for a test to
initialize multiple VMs with different seeds.

And as evidenced by memstress_guest_code(), it's also a-ok to instantiate
more RNGs using the global seed (or a modified version of it).  The goal
of the global RNG is purely to ensure that _a_ source of random numbers is
available, it doesn't have to be the _only_ RNG.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 .../testing/selftests/kvm/dirty_log_perf_test.c |  9 ++++-----
 tools/testing/selftests/kvm/dirty_log_test.c    | 17 +----------------
 tools/testing/selftests/kvm/include/memstress.h |  1 -
 tools/testing/selftests/kvm/include/test_util.h |  8 ++++++++
 tools/testing/selftests/kvm/lib/kvm_util.c      |  9 +++++++++
 tools/testing/selftests/kvm/lib/memstress.c     |  8 +-------
 6 files changed, 23 insertions(+), 29 deletions(-)

diff --git a/tools/testing/selftests/kvm/dirty_log_perf_test.c b/tools/testing/selftests/kvm/dirty_log_perf_test.c
index 504f6fe980e8..5cda9780c378 100644
--- a/tools/testing/selftests/kvm/dirty_log_perf_test.c
+++ b/tools/testing/selftests/kvm/dirty_log_perf_test.c
@@ -132,7 +132,6 @@ struct test_params {
 	enum vm_mem_backing_src_type backing_src;
 	int slots;
 	uint32_t write_percent;
-	uint32_t random_seed;
 	bool random_access;
 };
 
@@ -156,8 +155,6 @@ static void run_test(enum vm_guest_mode mode, void *arg)
 				 p->slots, p->backing_src,
 				 p->partition_vcpu_memory_access);
 
-	pr_info("Random seed: %u\n", p->random_seed);
-	memstress_set_random_seed(vm, p->random_seed);
 	memstress_set_write_percent(vm, p->write_percent);
 
 	guest_num_pages = (nr_vcpus * guest_percpu_mem_size) >> vm->page_shift;
@@ -346,11 +343,13 @@ int main(int argc, char *argv[])
 		.partition_vcpu_memory_access = true,
 		.backing_src = DEFAULT_VM_MEM_SRC,
 		.slots = 1,
-		.random_seed = 1,
 		.write_percent = 100,
 	};
 	int opt;
 
+	/* Override the seed to be deterministic by default. */
+	guest_random_seed = 1;
+
 	dirty_log_manual_caps =
 		kvm_check_cap(KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2);
 	dirty_log_manual_caps &= (KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE |
@@ -395,7 +394,7 @@ int main(int argc, char *argv[])
 			p.phys_offset = strtoull(optarg, NULL, 0);
 			break;
 		case 'r':
-			p.random_seed = atoi_positive("Random seed", optarg);
+			guest_random_seed = atoi_positive("Random seed", optarg);
 			break;
 		case 's':
 			p.backing_src = parse_backing_src_type(optarg);
diff --git a/tools/testing/selftests/kvm/dirty_log_test.c b/tools/testing/selftests/kvm/dirty_log_test.c
index eaad5b20854c..525d92dda646 100644
--- a/tools/testing/selftests/kvm/dirty_log_test.c
+++ b/tools/testing/selftests/kvm/dirty_log_test.c
@@ -76,7 +76,6 @@
 static uint64_t host_page_size;
 static uint64_t guest_page_size;
 static uint64_t guest_num_pages;
-static uint64_t random_array[TEST_PAGES_PER_LOOP];
 static uint64_t iteration;
 
 /*
@@ -115,13 +114,12 @@ static void guest_code(void)
 	while (true) {
 		for (i = 0; i < TEST_PAGES_PER_LOOP; i++) {
 			addr = guest_test_virt_mem;
-			addr += (READ_ONCE(random_array[i]) % guest_num_pages)
+			addr += (guest_random_u64(&guest_rng) % guest_num_pages)
 				* guest_page_size;
 			addr = align_down(addr, host_page_size);
 			*(uint64_t *)addr = READ_ONCE(iteration);
 		}
 
-		/* Tell the host that we need more random numbers */
 		GUEST_SYNC(1);
 	}
 }
@@ -508,20 +506,10 @@ static void log_mode_after_vcpu_run(struct kvm_vcpu *vcpu, int ret, int err)
 		mode->after_vcpu_run(vcpu, ret, err);
 }
 
-static void generate_random_array(uint64_t *guest_array, uint64_t size)
-{
-	uint64_t i;
-
-	for (i = 0; i < size; i++)
-		guest_array[i] = random();
-}
-
 static void *vcpu_worker(void *data)
 {
 	int ret;
 	struct kvm_vcpu *vcpu = data;
-	struct kvm_vm *vm = vcpu->vm;
-	uint64_t *guest_array;
 	uint64_t pages_count = 0;
 	struct kvm_signal_mask *sigmask = alloca(offsetof(struct kvm_signal_mask, sigset)
 						 + sizeof(sigset_t));
@@ -540,11 +528,8 @@ static void *vcpu_worker(void *data)
 	sigemptyset(sigset);
 	sigaddset(sigset, SIG_IPI);
 
-	guest_array = addr_gva2hva(vm, (vm_vaddr_t)random_array);
-
 	while (!READ_ONCE(host_quit)) {
 		/* Clear any existing kick signals */
-		generate_random_array(guest_array, TEST_PAGES_PER_LOOP);
 		pages_count += TEST_PAGES_PER_LOOP;
 		/* Let the guest dirty the random pages */
 		ret = __vcpu_run(vcpu);
diff --git a/tools/testing/selftests/kvm/include/memstress.h b/tools/testing/selftests/kvm/include/memstress.h
index ce4e603050ea..9071eb6dea60 100644
--- a/tools/testing/selftests/kvm/include/memstress.h
+++ b/tools/testing/selftests/kvm/include/memstress.h
@@ -62,7 +62,6 @@ struct kvm_vm *memstress_create_vm(enum vm_guest_mode mode, int nr_vcpus,
 void memstress_destroy_vm(struct kvm_vm *vm);
 
 void memstress_set_write_percent(struct kvm_vm *vm, uint32_t write_percent);
-void memstress_set_random_seed(struct kvm_vm *vm, uint32_t random_seed);
 void memstress_set_random_access(struct kvm_vm *vm, bool random_access);
 
 void memstress_start_vcpu_threads(int vcpus, void (*vcpu_fn)(struct memstress_vcpu_args *));
diff --git a/tools/testing/selftests/kvm/include/test_util.h b/tools/testing/selftests/kvm/include/test_util.h
index 8a6e30612c86..4b78fb7e539e 100644
--- a/tools/testing/selftests/kvm/include/test_util.h
+++ b/tools/testing/selftests/kvm/include/test_util.h
@@ -91,9 +91,17 @@ struct guest_random_state {
 	uint32_t seed;
 };
 
+extern uint32_t guest_random_seed;
+extern struct guest_random_state guest_rng;
+
 struct guest_random_state new_guest_random_state(uint32_t seed);
 uint32_t guest_random_u32(struct guest_random_state *state);
 
+static inline uint64_t guest_random_u64(struct guest_random_state *state)
+{
+	return ((uint64_t)guest_random_u32(state) << 32) | guest_random_u32(state);
+}
+
 enum vm_mem_backing_src_type {
 	VM_MEM_SRC_ANONYMOUS,
 	VM_MEM_SRC_ANONYMOUS_THP,
diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c
index b2262b5fad9e..02d7fa70d440 100644
--- a/tools/testing/selftests/kvm/lib/kvm_util.c
+++ b/tools/testing/selftests/kvm/lib/kvm_util.c
@@ -20,6 +20,9 @@
 
 #define KVM_UTIL_MIN_PFN	2
 
+uint32_t guest_random_seed;
+struct guest_random_state guest_rng;
+
 static int vcpu_mmap_sz(void);
 
 int open_path_or_exit(const char *path, int flags)
@@ -433,6 +436,10 @@ struct kvm_vm *__vm_create(struct vm_shape shape, uint32_t nr_runnable_vcpus,
 	slot0 = memslot2region(vm, 0);
 	ucall_init(vm, slot0->region.guest_phys_addr + slot0->region.memory_size);
 
+	pr_info("Random seed: 0x%x\n", guest_random_seed);
+	guest_rng = new_guest_random_state(guest_random_seed);
+	sync_global_to_guest(vm, guest_rng);
+
 	kvm_arch_vm_post_create(vm);
 
 	return vm;
@@ -2306,6 +2313,8 @@ void __attribute((constructor)) kvm_selftest_init(void)
 	/* Tell stdout not to buffer its content. */
 	setbuf(stdout, NULL);
 
+	guest_random_seed = random();
+
 	kvm_selftest_arch_init();
 }
 
diff --git a/tools/testing/selftests/kvm/lib/memstress.c b/tools/testing/selftests/kvm/lib/memstress.c
index cf2c73971308..f8bfb4988368 100644
--- a/tools/testing/selftests/kvm/lib/memstress.c
+++ b/tools/testing/selftests/kvm/lib/memstress.c
@@ -56,7 +56,7 @@ void memstress_guest_code(uint32_t vcpu_idx)
 	uint64_t page;
 	int i;
 
-	rand_state = new_guest_random_state(args->random_seed + vcpu_idx);
+	rand_state = new_guest_random_state(guest_random_seed + vcpu_idx);
 
 	gva = vcpu_args->gva;
 	pages = vcpu_args->pages;
@@ -243,12 +243,6 @@ void memstress_set_write_percent(struct kvm_vm *vm, uint32_t write_percent)
 	sync_global_to_guest(vm, memstress_args.write_percent);
 }
 
-void memstress_set_random_seed(struct kvm_vm *vm, uint32_t random_seed)
-{
-	memstress_args.random_seed = random_seed;
-	sync_global_to_guest(vm, memstress_args.random_seed);
-}
-
 void memstress_set_random_access(struct kvm_vm *vm, bool random_access)
 {
 	memstress_args.random_access = random_access;
-- 
2.44.0.291.gc1ea87d7ee-goog

[PATCH 2/5] KVM: selftests: Provide an API for getting a random bool from an RNG

[Sean Christopherson]

Move memstress' random bool logic into common code to avoid reinventing
the wheel for basic yes/no decisions.  Provide an outer wrapper to handle
the basic/common case of just wanting a 50/50 chance of something
happening.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 tools/testing/selftests/kvm/include/test_util.h | 11 +++++++++++
 tools/testing/selftests/kvm/lib/memstress.c     |  2 +-
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/kvm/include/test_util.h b/tools/testing/selftests/kvm/include/test_util.h
index 4b78fb7e539e..3e473058849f 100644
--- a/tools/testing/selftests/kvm/include/test_util.h
+++ b/tools/testing/selftests/kvm/include/test_util.h
@@ -97,6 +97,17 @@ extern struct guest_random_state guest_rng;
 struct guest_random_state new_guest_random_state(uint32_t seed);
 uint32_t guest_random_u32(struct guest_random_state *state);
 
+static inline bool __guest_random_bool(struct guest_random_state *state,
+				       uint8_t percent)
+{
+	return (guest_random_u32(state) % 100) < percent;
+}
+
+static inline bool guest_random_bool(struct guest_random_state *state)
+{
+	return __guest_random_bool(state, 50);
+}
+
 static inline uint64_t guest_random_u64(struct guest_random_state *state)
 {
 	return ((uint64_t)guest_random_u32(state) << 32) | guest_random_u32(state);
diff --git a/tools/testing/selftests/kvm/lib/memstress.c b/tools/testing/selftests/kvm/lib/memstress.c
index f8bfb4988368..2d49a5643b71 100644
--- a/tools/testing/selftests/kvm/lib/memstress.c
+++ b/tools/testing/selftests/kvm/lib/memstress.c
@@ -76,7 +76,7 @@ void memstress_guest_code(uint32_t vcpu_idx)
 
 			addr = gva + (page * args->guest_page_size);
 
-			if (guest_random_u32(&rand_state) % 100 < args->write_percent)
+			if (__guest_random_bool(&rand_state, args->write_percent))
 				*(uint64_t *)addr = 0x0123456789ABCDEF;
 			else
 				READ_ONCE(*(uint64_t *)addr);
-- 
2.44.0.291.gc1ea87d7ee-goog

[PATCH 3/5] KVM: selftests: Add global snapshot of kvm_is_forced_emulation_enabled()

[Sean Christopherson]

Add a global snapshot of kvm_is_forced_emulation_enabled() and sync it to
all VMs by default so that core library code can force emulation, e.g. to
allow for easier testing of the intersections between emulation and other
features in KVM.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 .../selftests/kvm/include/x86_64/kvm_util_arch.h       |  2 ++
 tools/testing/selftests/kvm/lib/x86_64/processor.c     |  3 +++
 tools/testing/selftests/kvm/x86_64/pmu_counters_test.c |  3 ---
 .../selftests/kvm/x86_64/userspace_msr_exit_test.c     | 10 ++--------
 4 files changed, 7 insertions(+), 11 deletions(-)

diff --git a/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h b/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h
index 9f1725192aa2..41aba476640a 100644
--- a/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h
+++ b/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h
@@ -5,6 +5,8 @@
 #include <stdbool.h>
 #include <stdint.h>
 
+extern bool is_forced_emulation_enabled;
+
 struct kvm_vm_arch {
 	uint64_t c_bit;
 	uint64_t s_bit;
diff --git a/tools/testing/selftests/kvm/lib/x86_64/processor.c b/tools/testing/selftests/kvm/lib/x86_64/processor.c
index 74a4c736c9ae..452d092ae050 100644
--- a/tools/testing/selftests/kvm/lib/x86_64/processor.c
+++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c
@@ -23,6 +23,7 @@
 vm_vaddr_t exception_handlers;
 bool host_cpu_is_amd;
 bool host_cpu_is_intel;
+bool is_forced_emulation_enabled;
 
 static void regs_dump(FILE *stream, struct kvm_regs *regs, uint8_t indent)
 {
@@ -577,6 +578,7 @@ void kvm_arch_vm_post_create(struct kvm_vm *vm)
 	vm_create_irqchip(vm);
 	sync_global_to_guest(vm, host_cpu_is_intel);
 	sync_global_to_guest(vm, host_cpu_is_amd);
+	sync_global_to_guest(vm, is_forced_emulation_enabled);
 
 	if (vm->subtype == VM_SUBTYPE_SEV)
 		sev_vm_init(vm);
@@ -1344,6 +1346,7 @@ void kvm_selftest_arch_init(void)
 {
 	host_cpu_is_intel = this_cpu_is_intel();
 	host_cpu_is_amd = this_cpu_is_amd();
+	is_forced_emulation_enabled = kvm_is_forced_emulation_enabled();
 }
 
 bool sys_clocksource_is_based_on_tsc(void)
diff --git a/tools/testing/selftests/kvm/x86_64/pmu_counters_test.c b/tools/testing/selftests/kvm/x86_64/pmu_counters_test.c
index 29609b52f8fa..7f7d4348e85c 100644
--- a/tools/testing/selftests/kvm/x86_64/pmu_counters_test.c
+++ b/tools/testing/selftests/kvm/x86_64/pmu_counters_test.c
@@ -21,7 +21,6 @@
 
 static uint8_t kvm_pmu_version;
 static bool kvm_has_perf_caps;
-static bool is_forced_emulation_enabled;
 
 static struct kvm_vm *pmu_vm_create_with_one_vcpu(struct kvm_vcpu **vcpu,
 						  void *guest_code,
@@ -35,7 +34,6 @@ static struct kvm_vm *pmu_vm_create_with_one_vcpu(struct kvm_vcpu **vcpu,
 	vcpu_init_descriptor_tables(*vcpu);
 
 	sync_global_to_guest(vm, kvm_pmu_version);
-	sync_global_to_guest(vm, is_forced_emulation_enabled);
 
 	/*
 	 * Set PERF_CAPABILITIES before PMU version as KVM disallows enabling
@@ -612,7 +610,6 @@ int main(int argc, char *argv[])
 
 	kvm_pmu_version = kvm_cpu_property(X86_PROPERTY_PMU_VERSION);
 	kvm_has_perf_caps = kvm_cpu_has(X86_FEATURE_PDCM);
-	is_forced_emulation_enabled = kvm_is_forced_emulation_enabled();
 
 	test_intel_counters();
 
diff --git a/tools/testing/selftests/kvm/x86_64/userspace_msr_exit_test.c b/tools/testing/selftests/kvm/x86_64/userspace_msr_exit_test.c
index f4f61a2d2464..69917bde69dc 100644
--- a/tools/testing/selftests/kvm/x86_64/userspace_msr_exit_test.c
+++ b/tools/testing/selftests/kvm/x86_64/userspace_msr_exit_test.c
@@ -13,8 +13,6 @@
 #include "kvm_util.h"
 #include "vmx.h"
 
-static bool fep_available;
-
 #define MSR_NON_EXISTENT 0x474f4f00
 
 static u64 deny_bits = 0;
@@ -258,7 +256,7 @@ static void guest_code_filter_allow(void)
 	GUEST_ASSERT(data == 2);
 	GUEST_ASSERT(guest_exception_count == 0);
 
-	if (fep_available) {
+	if (is_forced_emulation_enabled) {
 		/* Let userspace know we aren't done. */
 		GUEST_SYNC(0);
 
@@ -520,8 +518,6 @@ KVM_ONE_VCPU_TEST(user_msr, msr_filter_allow, guest_code_filter_allow)
 	uint64_t cmd;
 	int rc;
 
-	sync_global_to_guest(vm, fep_available);
-
 	rc = kvm_check_cap(KVM_CAP_X86_USER_SPACE_MSR);
 	TEST_ASSERT(rc, "KVM_CAP_X86_USER_SPACE_MSR is available");
 	vm_enable_cap(vm, KVM_CAP_X86_USER_SPACE_MSR, KVM_MSR_EXIT_REASON_FILTER);
@@ -551,7 +547,7 @@ KVM_ONE_VCPU_TEST(user_msr, msr_filter_allow, guest_code_filter_allow)
 	vcpu_run(vcpu);
 	cmd = process_ucall(vcpu);
 
-	if (fep_available) {
+	if (is_forced_emulation_enabled) {
 		TEST_ASSERT_EQ(cmd, UCALL_SYNC);
 		vm_install_exception_handler(vm, GP_VECTOR, guest_fep_gp_handler);
 
@@ -774,7 +770,5 @@ KVM_ONE_VCPU_TEST(user_msr, user_exit_msr_flags, NULL)
 
 int main(int argc, char *argv[])
 {
-	fep_available = kvm_is_forced_emulation_enabled();
-
 	return test_harness_run(argc, argv);
 }
-- 
2.44.0.291.gc1ea87d7ee-goog

[PATCH 4/5] KVM: selftests: Add vcpu_arch_put_guest() to do writes from guest code

[Sean Christopherson]

Introduce a macro, vcpu_arch_put_guest(), for "putting" values to memory
from guest code in "interesting" situations, e.g. when writing memory that
is being dirty logged.  Structure the macro so that arch code can provide
a custom implementation, e.g. x86 will use the macro to force emulation of
the access.

Use the helper in dirty_log_test, which is of particular interest (see
above), and in xen_shinfo_test, which isn't all that interesting, but
provides a second usage of the macro with a different size operand
(uint8_t versus uint64_t), i.e. to help verify that the macro works for
more than just 64-bit values.

Use "put" as the verb to align with the kernel's {get,put}_user()
terminology.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 tools/testing/selftests/kvm/dirty_log_test.c         | 5 +++--
 tools/testing/selftests/kvm/include/kvm_util_base.h  | 3 +++
 tools/testing/selftests/kvm/x86_64/xen_shinfo_test.c | 5 +++--
 3 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/tools/testing/selftests/kvm/dirty_log_test.c b/tools/testing/selftests/kvm/dirty_log_test.c
index 525d92dda646..e3f67f4584fb 100644
--- a/tools/testing/selftests/kvm/dirty_log_test.c
+++ b/tools/testing/selftests/kvm/dirty_log_test.c
@@ -108,7 +108,7 @@ static void guest_code(void)
 	 */
 	for (i = 0; i < guest_num_pages; i++) {
 		addr = guest_test_virt_mem + i * guest_page_size;
-		*(uint64_t *)addr = READ_ONCE(iteration);
+		vcpu_arch_put_guest(*(uint64_t *)addr, READ_ONCE(iteration));
 	}
 
 	while (true) {
@@ -117,7 +117,8 @@ static void guest_code(void)
 			addr += (guest_random_u64(&guest_rng) % guest_num_pages)
 				* guest_page_size;
 			addr = align_down(addr, host_page_size);
-			*(uint64_t *)addr = READ_ONCE(iteration);
+
+			vcpu_arch_put_guest(*(uint64_t *)addr, READ_ONCE(iteration));
 		}
 
 		GUEST_SYNC(1);
diff --git a/tools/testing/selftests/kvm/include/kvm_util_base.h b/tools/testing/selftests/kvm/include/kvm_util_base.h
index 3e0db283a46a..9110efa1bc12 100644
--- a/tools/testing/selftests/kvm/include/kvm_util_base.h
+++ b/tools/testing/selftests/kvm/include/kvm_util_base.h
@@ -610,6 +610,9 @@ void *addr_gva2hva(struct kvm_vm *vm, vm_vaddr_t gva);
 vm_paddr_t addr_hva2gpa(struct kvm_vm *vm, void *hva);
 void *addr_gpa2alias(struct kvm_vm *vm, vm_paddr_t gpa);
 
+#ifndef vcpu_arch_put_guest
+#define vcpu_arch_put_guest(mem, val) do { (mem) = (val); } while (0)
+#endif
 
 static inline vm_paddr_t vm_untag_gpa(struct kvm_vm *vm, vm_paddr_t gpa)
 {
diff --git a/tools/testing/selftests/kvm/x86_64/xen_shinfo_test.c b/tools/testing/selftests/kvm/x86_64/xen_shinfo_test.c
index d2ea0435f4f7..1ba06551526b 100644
--- a/tools/testing/selftests/kvm/x86_64/xen_shinfo_test.c
+++ b/tools/testing/selftests/kvm/x86_64/xen_shinfo_test.c
@@ -171,8 +171,9 @@ static volatile bool guest_saw_irq;
 static void evtchn_handler(struct ex_regs *regs)
 {
 	struct vcpu_info *vi = (void *)VCPU_INFO_VADDR;
-	vi->evtchn_upcall_pending = 0;
-	vi->evtchn_pending_sel = 0;
+
+	vcpu_arch_put_guest(vi->evtchn_upcall_pending, 0);
+	vcpu_arch_put_guest(vi->evtchn_pending_sel, 0);
 	guest_saw_irq = true;
 
 	GUEST_SYNC(TEST_GUEST_SAW_IRQ);
-- 
2.44.0.291.gc1ea87d7ee-goog

[PATCH 5/5] KVM: selftests: Randomly force emulation on x86 writes from guest code

[Sean Christopherson]

Override vcpu_arch_put_guest() to randomly force emulation on supported
accesses.  Force emulation of LOCK CMPXCHG as well as a regular MOV to
stress KVM's emulation of atomic accesses, which has a unique path in
KVM's emulator.

Arbitrarily give all the decisions 50/50 odds; absent much, much more
sophisticated infrastructure for generating random numbers, it's highly
unlikely that doing more than a coin flip with affect selftests' ability
to find KVM bugs.

This is effectively a regression test for commit 910c57dfa4d1 ("KVM: x86:
Mark target gfn of emulated atomic instruction as dirty").

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 .../kvm/include/x86_64/kvm_util_arch.h        | 21 +++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h b/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h
index 41aba476640a..d0b587c38e07 100644
--- a/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h
+++ b/tools/testing/selftests/kvm/include/x86_64/kvm_util_arch.h
@@ -5,6 +5,8 @@
 #include <stdbool.h>
 #include <stdint.h>
 
+#include "test_util.h"
+
 extern bool is_forced_emulation_enabled;
 
 struct kvm_vm_arch {
@@ -22,4 +24,23 @@ static inline bool __vm_arch_has_protected_memory(struct kvm_vm_arch *arch)
 #define vm_arch_has_protected_memory(vm) \
 	__vm_arch_has_protected_memory(&(vm)->arch)
 
+#define vcpu_arch_put_guest(mem, __val)							\
+do {											\
+	const typeof(mem) val = (__val);						\
+											\
+	if (!is_forced_emulation_enabled || guest_random_bool(&guest_rng)) {		\
+		(mem) = val;								\
+	} else if (guest_random_bool(&guest_rng)) {					\
+		__asm__ __volatile__(KVM_FEP "mov %1, %0"				\
+				     : "+m" (mem)					\
+				     : "r" (val) : "memory");				\
+	} else {									\
+		uint64_t __old = READ_ONCE(mem);					\
+											\
+		__asm__ __volatile__(KVM_FEP LOCK_PREFIX "cmpxchg %[new], %[ptr]"	\
+				     : [ptr] "+m" (mem), [old] "+a" (__old)		\
+				     : [new]"r" (val) : "memory", "cc");		\
+	}										\
+} while (0)
+
 #endif  // SELFTEST_KVM_UTIL_ARCH_H
-- 
2.44.0.291.gc1ea87d7ee-goog

Re: [PATCH 0/5] KVM: selftests: Introduce vcpu_arch_put_guest()

[Sean Christopherson]

On Thu, 14 Mar 2024 11:54:53 -0700, Sean Christopherson wrote:
> The end goal of this series is to add a regression test for commit
> 910c57dfa4d1 ("KVM: x86: Mark target gfn of emulated atomic instruction
> as dirty"), *without* polluting the common dirty_log_test.c code with
> gory x86 details.
> 
> The regression test requires forcing KVM to emulate a guest atomic RMW
> access, which is done via a magic instruction prefix/opcode that is
> guarded by an off-by-default KVM module param.
> 
> [...]

Applied to kvm-x86 selftests_utils, thanks!

[1/5] KVM: selftests: Provide a global pseudo-RNG instance for all tests
      https://github.com/kvm-x86/linux/commit/cb6c6914788f
[2/5] KVM: selftests: Provide an API for getting a random bool from an RNG
      https://github.com/kvm-x86/linux/commit/73369acd9fbd
[3/5] KVM: selftests: Add global snapshot of kvm_is_forced_emulation_enabled()
      https://github.com/kvm-x86/linux/commit/e1ff11525d3c
[4/5] KVM: selftests: Add vcpu_arch_put_guest() to do writes from guest code
      https://github.com/kvm-x86/linux/commit/2f2bc6af6aa8
[5/5] KVM: selftests: Randomly force emulation on x86 writes from guest code
      https://github.com/kvm-x86/linux/commit/87aa264cd89d

--
https://github.com/kvm-x86/linux/tree/next