Re: [PATCH v2 4/5] KVM: selftests: aarch64: Introduce pmu_event_filter_test
From: Eric Auger
Date: Thu Dec 14 2023 - 14:29:43 EST
Hi Shaoqin,
On 12/14/23 14:45, Eric Auger wrote:
> Hi Shaoqin,
>
> On 11/29/23 08:27, Shaoqin Huang wrote:
>> Introduce pmu_event_filter_test for arm64 platforms. The test configures
>> PMUv3 for a vCPU, and sets different pmu event filters for the vCPU, and
>> check if the guest can use those events which user allow and can't use
>> those events which use deny.
>>
>> This test refactor the create_vpmu_vm() and make it a wrapper for
>> __create_vpmu_vm(), which allows some extra init code before
>> KVM_ARM_VCPU_PMU_V3_INIT.
>>
>> And this test use the KVM_ARM_VCPU_PMU_V3_FILTER attribute to set the
>> pmu event filter in KVM. And choose to filter two common event
>> branches_retired and instructions_retired, and let guest use the two
>> events in pmu. And check if the result is expected.
>>
>> Signed-off-by: Shaoqin Huang <shahuang@xxxxxxxxxx>
>> ---
>> tools/testing/selftests/kvm/Makefile | 1 +
>> .../kvm/aarch64/pmu_event_filter_test.c | 231 ++++++++++++++++++
>> .../selftests/kvm/include/aarch64/vpmu.h | 4 +
>> .../testing/selftests/kvm/lib/aarch64/vpmu.c | 14 +-
>> 4 files changed, 248 insertions(+), 2 deletions(-)
>> create mode 100644 tools/testing/selftests/kvm/aarch64/pmu_event_filter_test.c
>>
>> diff --git a/tools/testing/selftests/kvm/Makefile b/tools/testing/selftests/kvm/Makefile
>> index b60852c222ac..5f126e1a1dbf 100644
>> --- a/tools/testing/selftests/kvm/Makefile
>> +++ b/tools/testing/selftests/kvm/Makefile
>> @@ -155,6 +155,7 @@ TEST_GEN_PROGS_aarch64 += aarch64/arch_timer
>> TEST_GEN_PROGS_aarch64 += aarch64/debug-exceptions
>> TEST_GEN_PROGS_aarch64 += aarch64/hypercalls
>> TEST_GEN_PROGS_aarch64 += aarch64/page_fault_test
>> +TEST_GEN_PROGS_aarch64 += aarch64/pmu_event_filter_test
>> TEST_GEN_PROGS_aarch64 += aarch64/psci_test
>> TEST_GEN_PROGS_aarch64 += aarch64/set_id_regs
>> TEST_GEN_PROGS_aarch64 += aarch64/smccc_filter
>> diff --git a/tools/testing/selftests/kvm/aarch64/pmu_event_filter_test.c b/tools/testing/selftests/kvm/aarch64/pmu_event_filter_test.c
>> new file mode 100644
>> index 000000000000..0e652fbdb37a
>> --- /dev/null
>> +++ b/tools/testing/selftests/kvm/aarch64/pmu_event_filter_test.c
>> @@ -0,0 +1,231 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +/*
>> + * pmu_event_filter_test - Test user limit pmu event for guest.
>> + *
>> + * Copyright (c) 2023 Red Hat, Inc.
>> + *
>> + * This test checks if the guest only see the limited pmu event that userspace
>> + * sets, if the guest can use those events which user allow, and if the guest
>> + * can't use those events which user deny.
>> + * This test runs only when KVM_CAP_ARM_PMU_V3, KVM_ARM_VCPU_PMU_V3_FILTER
>> + * is supported on the host.
>> + */
>> +#include <kvm_util.h>
>> +#include <processor.h>
>> +#include <vgic.h>
>> +#include <vpmu.h>
>> +#include <test_util.h>
>> +#include <perf/arm_pmuv3.h>
>> +
>> +struct {
>> + uint64_t branches_retired;
>> + uint64_t instructions_retired;
>> +} pmc_results;
>> +
>> +static struct vpmu_vm *vpmu_vm;
>> +static uint64_t pmceid0;
>> +
>> +#define FILTER_NR 10
>> +
>> +struct test_desc {
>> + const char *name;
>> + void (*check_result)(void);
>> + struct kvm_pmu_event_filter filter[FILTER_NR];
>> +};
>> +> +#define __DEFINE_FILTER(base, num, act) \
>> + ((struct kvm_pmu_event_filter) { \
>> + .base_event = base, \
>> + .nevents = num, \
>> + .action = act, \
>> + })
>> +
>> +#define DEFINE_FILTER(base, act) __DEFINE_FILTER(base, 1, act)
>> +
>> +#define EMPTY_FILTER { 0 }
>> +
>> +#define SW_INCR 0x0
>> +#define INST_RETIRED 0x8
>> +#define BR_RETIRED 0x21
>> +
>> +#define NUM_BRANCHES 10
>> +
>> +static void run_and_measure_loop(void)
>> +{
>> + asm volatile(
>> + " mov x10, %[loop]\n"
>> + "1: sub x10, x10, #1\n"
>> + " cmp x10, #0x0\n"
>> + " b.gt 1b\n"
>> + :
>> + : [loop] "r" (NUM_BRANCHES)
>> + : "x10", "cc");
>> +}
>> +
>> +static void guest_code(void)
>> +{
>> + uint64_t pmcr = read_sysreg(pmcr_el0);
>> +
>> + pmu_disable_reset();
>> +
>> + write_pmevtypern(0, BR_RETIRED);
>> + write_pmevtypern(1, INST_RETIRED);
>> + enable_counter(0);
>> + enable_counter(1);
>> + write_sysreg(pmcr | ARMV8_PMU_PMCR_E, pmcr_el0);
>> +
>> + run_and_measure_loop();
>> +
>> + write_sysreg(pmcr, pmcr_el0);
>> +
>> + pmc_results.branches_retired = read_sysreg(pmevcntr0_el0);
>> + pmc_results.instructions_retired = read_sysreg(pmevcntr1_el0);
>> +
>> + GUEST_DONE();
>> +}
>> +
>> +static void guest_get_pmceid0(void)
>> +{
>> + uint64_t pmceid0 = read_sysreg(pmceid0_el0);
>> +
>> + GUEST_PRINTF("%lx\n", pmceid0);
>> +
>> + GUEST_DONE();
>> +}
>> +
>> +static void pmu_event_filter_init(struct vpmu_vm *vm, void *arg)
>> +{
>> + struct kvm_device_attr attr = {
>> + .group = KVM_ARM_VCPU_PMU_V3_CTRL,
>> + .attr = KVM_ARM_VCPU_PMU_V3_FILTER,
>> + };
>> + struct kvm_pmu_event_filter *filter = (struct kvm_pmu_event_filter *)arg;
>> +
>> + while (filter && filter->nevents != 0) {
>> + attr.addr = (uint64_t)filter;
>> + vcpu_ioctl(vm->vcpu, KVM_SET_DEVICE_ATTR, &attr);
>> + filter++;
>> + }
>> +}
>> +
>> +static void create_vpmu_vm_with_filter(void *guest_code,
>> + struct kvm_pmu_event_filter *filter)
>> +{
>> + vpmu_vm = __create_vpmu_vm(guest_code, pmu_event_filter_init, filter);
>> +}
>> +
>> +static void run_vcpu(struct kvm_vcpu *vcpu)
>> +{
>> + struct ucall uc;
>> +
>> + while (1) {
>> + vcpu_run(vcpu);
>> + switch (get_ucall(vcpu, &uc)) {
>> + case UCALL_DONE:
>> + return;
>> + case UCALL_PRINTF:
>> + pmceid0 = strtoll(uc.buffer, NULL, 16);
>> + break;
>> + default:
>> + TEST_FAIL("Unknown ucall %lu", uc.cmd);
>> + }
>> + }
>> +}
>> +
>> +static void check_pmc_counting(void)
>> +{
>> + uint64_t br = pmc_results.branches_retired;
>> + uint64_t ir = pmc_results.instructions_retired;
>> +
>> + TEST_ASSERT(br && br == NUM_BRANCHES, "Branch instructions retired = "
>> + "%lu (expected %u)", br, NUM_BRANCHES);
> have you tested on several machines? My experience with some events
> (MEM_ACCESS for instance) is that you have variance (sometimes
> significant) on some event count. I am a little bit scared that having
> this br == NUM_BRANCHES check without taking into account some margin
> will cause failures on some HW.
I confirm the usual suspect, Amberwing, does not like this check ;-)
augere@qualcomm-amberwing-rep-06:~/UPSTREAM/linux/tools/testing/selftests/kvm/aarch64#
./pmu_event_filter_test
==== Test Assertion Failure ====
aarch64/pmu_event_filter_test.c:141: br && br == NUM_BRANCHES
pid=7750 tid=7750 errno=4 - Interrupted system call
1 0x0000000000401d6b: check_pmc_counting at pmu_event_filter_test.c:141
2 0x0000000000401967: run_test at pmu_event_filter_test.c:173
3 (inlined by) for_each_test at pmu_event_filter_test.c:198
4 (inlined by) main at pmu_event_filter_test.c:264
5 0x0000ffffaaa6c79b: ?? ??:0
6 0x0000ffffaaa6c86b: ?? ??:0
7 0x0000000000401aaf: _start at ??:?
Branch instructions retired = 15 (expected 10)
Eric
>
> in v1 I suggested to read to PMCEID* in a guest code to check if the
> event is supported. This method would also have the benefice to allow
> testing more complex filter range combinations.
>> + TEST_ASSERT(ir, "Instructions retired = %lu (expected > 0)", ir);
>> +}
>> +
>> +static void check_pmc_not_counting(void)
>> +{
>> + uint64_t br = pmc_results.branches_retired;
>> + uint64_t ir = pmc_results.instructions_retired;
>> +
>> + TEST_ASSERT(!br, "Branch instructions retired = %lu (expected 0)", br);
>> + TEST_ASSERT(!ir, "Instructions retired = %lu (expected 0)", ir);
>> +}
>> +
>> +static void run_vcpu_and_sync_pmc_results(void)
>> +{
>> + memset(&pmc_results, 0, sizeof(pmc_results));
>> + sync_global_to_guest(vpmu_vm->vm, pmc_results);
>> +
>> + run_vcpu(vpmu_vm->vcpu);
>> +
>> + sync_global_from_guest(vpmu_vm->vm, pmc_results);
>> +}
>> +
>> +static void run_test(struct test_desc *t)
>> +{
>> + pr_debug("Test: %s\n", t->name);
>> +
>> + create_vpmu_vm_with_filter(guest_code, t->filter);
>> +
>> + run_vcpu_and_sync_pmc_results();
>> +
>> + t->check_result();
>> +
>> + destroy_vpmu_vm(vpmu_vm);
>> +}
>> +
>> +static struct test_desc tests[] = {
>> + {"without_filter", check_pmc_counting, { EMPTY_FILTER }},
>> + {"member_allow_filter", check_pmc_counting,
>> + {DEFINE_FILTER(SW_INCR, 0), DEFINE_FILTER(INST_RETIRED, 0),
> Note the doc says that Event 0 (SW_INCR) is never filtered, as it
> doesn't count a hardware event
>
>
> I would use the defines exposed in the uapi
>> +#define KVM_PMU_EVENT_ALLOW 0
>> +#define KVM_PMU_EVENT_DENY 1
>> + DEFINE_FILTER(BR_RETIRED, 0), EMPTY_FILTER}},
>> + {"member_deny_filter", check_pmc_not_counting,
>> + {DEFINE_FILTER(SW_INCR, 1), DEFINE_FILTER(INST_RETIRED, 1),
> what is the purpose of SW_INCR. YOu do not seem to test it anyway?
>> + DEFINE_FILTER(BR_RETIRED, 1), EMPTY_FILTER}},
>> + {"not_member_deny_filter", check_pmc_counting,
>> + {DEFINE_FILTER(SW_INCR, 1), EMPTY_FILTER}},
>> + {"not_member_allow_filter", check_pmc_not_counting,
>> + {DEFINE_FILTER(SW_INCR, 0), EMPTY_FILTER}},
>> + { 0 }
>> +};
>> +
>> +static void for_each_test(void)
>> +{
>> + struct test_desc *t;
>> +
>> + for (t = &tests[0]; t->name; t++)
>> + run_test(t);
>> +}
>> +
>> +static bool kvm_supports_pmu_event_filter(void)
>> +{
>> + int r;
>> +
>> + vpmu_vm = create_vpmu_vm(guest_code);
>> +
>> + r = __kvm_has_device_attr(vpmu_vm->vcpu->fd, KVM_ARM_VCPU_PMU_V3_CTRL,
>> + KVM_ARM_VCPU_PMU_V3_FILTER);
> you can use __vcpu_has_device_attr directly
>> +
>> + destroy_vpmu_vm(vpmu_vm);
>> + return !r;
>> +}
>> +
>> +static bool host_pmu_supports_events(void)
>> +{
>> + vpmu_vm = create_vpmu_vm(guest_get_pmceid0);
>> +
>> + run_vcpu(vpmu_vm->vcpu);
>> +
>> + destroy_vpmu_vm(vpmu_vm);
>> +
>> + return pmceid0 & (BR_RETIRED | INST_RETIRED);
> this will return true if either event is supported. I suspect this is
> not what you want.
>> +}
>> +
>> +int main(void)
>> +{
>> + TEST_REQUIRE(kvm_has_cap(KVM_CAP_ARM_PMU_V3));
>> + TEST_REQUIRE(kvm_supports_pmu_event_filter());
>> + TEST_REQUIRE(host_pmu_supports_events());
>> +
>> + for_each_test();
>> +}
>> diff --git a/tools/testing/selftests/kvm/include/aarch64/vpmu.h b/tools/testing/selftests/kvm/include/aarch64/vpmu.h
>> index 644dae3814b5..f103d0824f8a 100644
>> --- a/tools/testing/selftests/kvm/include/aarch64/vpmu.h
>> +++ b/tools/testing/selftests/kvm/include/aarch64/vpmu.h
>> @@ -18,6 +18,10 @@ struct vpmu_vm {
>> int gic_fd;
>> };
>>
>> +struct vpmu_vm *__create_vpmu_vm(void *guest_code,
>> + void (*init_pmu)(struct vpmu_vm *vm, void *arg),
>> + void *arg);
>> +
>> struct vpmu_vm *create_vpmu_vm(void *guest_code);
>>
>> void destroy_vpmu_vm(struct vpmu_vm *vpmu_vm);
>> diff --git a/tools/testing/selftests/kvm/lib/aarch64/vpmu.c b/tools/testing/selftests/kvm/lib/aarch64/vpmu.c
>> index b3de8fdc555e..76ea03d607f1 100644
>> --- a/tools/testing/selftests/kvm/lib/aarch64/vpmu.c
>> +++ b/tools/testing/selftests/kvm/lib/aarch64/vpmu.c
>> @@ -7,8 +7,9 @@
>> #include <vpmu.h>
>> #include <perf/arm_pmuv3.h>
>>
>> -/* Create a VM that has one vCPU with PMUv3 configured. */
>> -struct vpmu_vm *create_vpmu_vm(void *guest_code)
>> +struct vpmu_vm *__create_vpmu_vm(void *guest_code,
>> + void (*init_pmu)(struct vpmu_vm *vm, void *arg),
>> + void *arg)
>> {
>> struct kvm_vcpu_init init;
>> uint8_t pmuver;
>> @@ -50,12 +51,21 @@ struct vpmu_vm *create_vpmu_vm(void *guest_code)
>> "Unexpected PMUVER (0x%x) on the vCPU with PMUv3", pmuver);
>>
>> /* Initialize vPMU */
>> + if (init_pmu)
>> + init_pmu(vpmu_vm, arg);
>> +
>> vcpu_ioctl(vpmu_vm->vcpu, KVM_SET_DEVICE_ATTR, &irq_attr);
>> vcpu_ioctl(vpmu_vm->vcpu, KVM_SET_DEVICE_ATTR, &init_attr);
>>
>> return vpmu_vm;
>> }
>>
>> +/* Create a VM that has one vCPU with PMUv3 configured. */
>> +struct vpmu_vm *create_vpmu_vm(void *guest_code)
>> +{
>> + return __create_vpmu_vm(guest_code, NULL, NULL);
>> +}
>> +
>> void destroy_vpmu_vm(struct vpmu_vm *vpmu_vm)
>> {
>> close(vpmu_vm->gic_fd);
>
> Thanks
>
> Eric