From c071ff41e1502990a0902c7c3fcf72a462e46330 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:00 +0000 Subject: KVM: selftests: Expose align() helpers to tests Refactor align() to work with non-pointers and split into separate helpers for aligning up vs. down. Add align_ptr_up() for use with pointers. Expose all helpers so that they can be used by tests and/or other utilities. The align_down() helper in particular will be used to ensure gpa alignment for hugepages. No functional change intended. [Added sepearate up/down helpers and replaced open-coded alignment bit math throughout the KVM selftests.] Signed-off-by: Sean Christopherson Signed-off-by: David Matlack Reviewed-by: Ben Gardon Message-Id: <20211111000310.1435032-3-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/lib/perf_test_util.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index 0ef80dbdc116..6b8d5020dc54 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -92,10 +92,10 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * perf_test_args.guest_page_size; - guest_test_phys_mem &= ~(perf_test_args.host_page_size - 1); + guest_test_phys_mem = align_down(guest_test_phys_mem, perf_test_args.host_page_size); #ifdef __s390x__ /* Align to 1M (segment size) */ - guest_test_phys_mem &= ~((1 << 20) - 1); + guest_test_phys_mem = align_down(guest_test_phys_mem, 1 << 20); #endif pr_info("guest physical test memory offset: 0x%lx\n", guest_test_phys_mem); -- cgit v1.2.3-70-g09d2 From 69cdcfa6f321da2cc1dd2e62fa4a9ee256299b18 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:02 +0000 Subject: KVM: selftests: Require GPA to be aligned when backed by hugepages Assert that the GPA for a memslot backed by a hugepage is aligned to the hugepage size and fix perf_test_util accordingly. Lack of GPA alignment prevents KVM from backing the guest with hugepages, e.g. x86's write-protection of hugepages when dirty logging is activated is otherwise not exercised. Add a comment explaining that guest_page_size is for non-huge pages to try and avoid confusion about what it actually tracks. Cc: Ben Gardon Cc: Yanan Wang Cc: Andrew Jones Cc: Peter Xu Cc: Aaron Lewis Signed-off-by: Sean Christopherson [Used get_backing_src_pagesz() to determine alignment dynamically.] Signed-off-by: David Matlack Message-Id: <20211111000310.1435032-5-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/lib/kvm_util.c | 2 ++ tools/testing/selftests/kvm/lib/perf_test_util.c | 7 ++++++- 2 files changed, 8 insertions(+), 1 deletion(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c index 07f37456bba0..1f6a01c33dce 100644 --- a/tools/testing/selftests/kvm/lib/kvm_util.c +++ b/tools/testing/selftests/kvm/lib/kvm_util.c @@ -875,6 +875,8 @@ void vm_userspace_mem_region_add(struct kvm_vm *vm, if (src_type == VM_MEM_SRC_ANONYMOUS_THP) alignment = max(backing_src_pagesz, alignment); + ASSERT_EQ(guest_paddr, align_up(guest_paddr, backing_src_pagesz)); + /* Add enough memory to align up if necessary */ if (alignment > 1) region->mmap_size += alignment; diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index 6b8d5020dc54..a015f267d945 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -55,11 +55,16 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, { struct kvm_vm *vm; uint64_t guest_num_pages; + uint64_t backing_src_pagesz = get_backing_src_pagesz(backing_src); int i; pr_info("Testing guest mode: %s\n", vm_guest_mode_string(mode)); perf_test_args.host_page_size = getpagesize(); + /* + * Snapshot the non-huge page size. This is used by the guest code to + * access/dirty pages at the logging granularity. + */ perf_test_args.guest_page_size = vm_guest_mode_params[mode].page_size; guest_num_pages = vm_adjust_num_guest_pages(mode, @@ -92,7 +97,7 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * perf_test_args.guest_page_size; - guest_test_phys_mem = align_down(guest_test_phys_mem, perf_test_args.host_page_size); + guest_test_phys_mem = align_down(guest_test_phys_mem, backing_src_pagesz); #ifdef __s390x__ /* Align to 1M (segment size) */ guest_test_phys_mem = align_down(guest_test_phys_mem, 1 << 20); -- cgit v1.2.3-70-g09d2 From b65e1051e489be4fe783cb14f1cd33235a0f9803 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:03 +0000 Subject: KVM: selftests: Use shorthand local var to access struct perf_tests_args Use 'pta' as a local pointer to the global perf_tests_args in order to shorten line lengths and make the code borderline readable. No functional change intended. Signed-off-by: Sean Christopherson Reviewed-by: Ben Gardon Signed-off-by: David Matlack Message-Id: <20211111000310.1435032-6-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/lib/perf_test_util.c | 35 +++++++++++++----------- 1 file changed, 19 insertions(+), 16 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index a015f267d945..ccdc950c829e 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -24,7 +24,8 @@ static uint64_t guest_test_virt_mem = DEFAULT_GUEST_TEST_MEM; */ static void guest_code(uint32_t vcpu_id) { - struct perf_test_vcpu_args *vcpu_args = &perf_test_args.vcpu_args[vcpu_id]; + struct perf_test_args *pta = &perf_test_args; + struct perf_test_vcpu_args *vcpu_args = &pta->vcpu_args[vcpu_id]; uint64_t gva; uint64_t pages; int i; @@ -37,9 +38,9 @@ static void guest_code(uint32_t vcpu_id) while (true) { for (i = 0; i < pages; i++) { - uint64_t addr = gva + (i * perf_test_args.guest_page_size); + uint64_t addr = gva + (i * pta->guest_page_size); - if (i % perf_test_args.wr_fract == 0) + if (i % pta->wr_fract == 0) *(uint64_t *)addr = 0x0123456789ABCDEF; else READ_ONCE(*(uint64_t *)addr); @@ -53,6 +54,7 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, uint64_t vcpu_memory_bytes, int slots, enum vm_mem_backing_src_type backing_src) { + struct perf_test_args *pta = &perf_test_args; struct kvm_vm *vm; uint64_t guest_num_pages; uint64_t backing_src_pagesz = get_backing_src_pagesz(backing_src); @@ -60,29 +62,29 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, pr_info("Testing guest mode: %s\n", vm_guest_mode_string(mode)); - perf_test_args.host_page_size = getpagesize(); + pta->host_page_size = getpagesize(); /* * Snapshot the non-huge page size. This is used by the guest code to * access/dirty pages at the logging granularity. */ - perf_test_args.guest_page_size = vm_guest_mode_params[mode].page_size; + pta->guest_page_size = vm_guest_mode_params[mode].page_size; guest_num_pages = vm_adjust_num_guest_pages(mode, - (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size); + (vcpus * vcpu_memory_bytes) / pta->guest_page_size); - TEST_ASSERT(vcpu_memory_bytes % perf_test_args.host_page_size == 0, + TEST_ASSERT(vcpu_memory_bytes % pta->host_page_size == 0, "Guest memory size is not host page size aligned."); - TEST_ASSERT(vcpu_memory_bytes % perf_test_args.guest_page_size == 0, + TEST_ASSERT(vcpu_memory_bytes % pta->guest_page_size == 0, "Guest memory size is not guest page size aligned."); TEST_ASSERT(guest_num_pages % slots == 0, "Guest memory cannot be evenly divided into %d slots.", slots); vm = vm_create_with_vcpus(mode, vcpus, DEFAULT_GUEST_PHY_PAGES, - (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size, + (vcpus * vcpu_memory_bytes) / pta->guest_page_size, 0, guest_code, NULL); - perf_test_args.vm = vm; + pta->vm = vm; /* * If there should be more memory in the guest test region than there @@ -96,7 +98,7 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, vcpu_memory_bytes); guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * - perf_test_args.guest_page_size; + pta->guest_page_size; guest_test_phys_mem = align_down(guest_test_phys_mem, backing_src_pagesz); #ifdef __s390x__ /* Align to 1M (segment size) */ @@ -108,7 +110,7 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, for (i = 0; i < slots; i++) { uint64_t region_pages = guest_num_pages / slots; vm_paddr_t region_start = guest_test_phys_mem + - region_pages * perf_test_args.guest_page_size * i; + region_pages * pta->guest_page_size * i; vm_userspace_mem_region_add(vm, backing_src, region_start, PERF_TEST_MEM_SLOT_INDEX + i, @@ -133,25 +135,26 @@ void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, uint64_t vcpu_memory_bytes, bool partition_vcpu_memory_access) { + struct perf_test_args *pta = &perf_test_args; vm_paddr_t vcpu_gpa; struct perf_test_vcpu_args *vcpu_args; int vcpu_id; for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { - vcpu_args = &perf_test_args.vcpu_args[vcpu_id]; + vcpu_args = &pta->vcpu_args[vcpu_id]; vcpu_args->vcpu_id = vcpu_id; if (partition_vcpu_memory_access) { vcpu_args->gva = guest_test_virt_mem + (vcpu_id * vcpu_memory_bytes); vcpu_args->pages = vcpu_memory_bytes / - perf_test_args.guest_page_size; + pta->guest_page_size; vcpu_gpa = guest_test_phys_mem + (vcpu_id * vcpu_memory_bytes); } else { vcpu_args->gva = guest_test_virt_mem; vcpu_args->pages = (vcpus * vcpu_memory_bytes) / - perf_test_args.guest_page_size; + pta->guest_page_size; vcpu_gpa = guest_test_phys_mem; } @@ -159,6 +162,6 @@ void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, pr_debug("Added VCPU %d with test mem gpa [%lx, %lx)\n", vcpu_id, vcpu_gpa, vcpu_gpa + - (vcpu_args->pages * perf_test_args.guest_page_size)); + (vcpu_args->pages * pta->guest_page_size)); } } -- cgit v1.2.3-70-g09d2 From 613d61182fffca6b36ea0df1e44927ccf45b1e9b Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:04 +0000 Subject: KVM: selftests: Capture per-vCPU GPA in perf_test_vcpu_args Capture the per-vCPU GPA in perf_test_vcpu_args so that tests can get the GPA without having to calculate the GPA on their own. No functional change intended. Signed-off-by: Sean Christopherson Reviewed-by: Ben Gardon Signed-off-by: David Matlack Message-Id: <20211111000310.1435032-7-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/include/perf_test_util.h | 1 + tools/testing/selftests/kvm/lib/perf_test_util.c | 9 ++++----- 2 files changed, 5 insertions(+), 5 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index df9f1a3a3ffb..20aec72fe7b8 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -18,6 +18,7 @@ #define PERF_TEST_MEM_SLOT_INDEX 1 struct perf_test_vcpu_args { + uint64_t gpa; uint64_t gva; uint64_t pages; diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index ccdc950c829e..d9c6bcb7964d 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -136,7 +136,6 @@ void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, bool partition_vcpu_memory_access) { struct perf_test_args *pta = &perf_test_args; - vm_paddr_t vcpu_gpa; struct perf_test_vcpu_args *vcpu_args; int vcpu_id; @@ -149,19 +148,19 @@ void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, (vcpu_id * vcpu_memory_bytes); vcpu_args->pages = vcpu_memory_bytes / pta->guest_page_size; - vcpu_gpa = guest_test_phys_mem + - (vcpu_id * vcpu_memory_bytes); + vcpu_args->gpa = guest_test_phys_mem + + (vcpu_id * vcpu_memory_bytes); } else { vcpu_args->gva = guest_test_virt_mem; vcpu_args->pages = (vcpus * vcpu_memory_bytes) / pta->guest_page_size; - vcpu_gpa = guest_test_phys_mem; + vcpu_args->gpa = guest_test_phys_mem; } vcpu_args_set(vm, vcpu_id, 1, vcpu_id); pr_debug("Added VCPU %d with test mem gpa [%lx, %lx)\n", - vcpu_id, vcpu_gpa, vcpu_gpa + + vcpu_id, vcpu_args->gpa, vcpu_args->gpa + (vcpu_args->pages * pta->guest_page_size)); } } -- cgit v1.2.3-70-g09d2 From b91b637f4a595c5be435e215f78b1a3bd8c252b3 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:06 +0000 Subject: KVM: selftests: Move per-VM GPA into perf_test_args Move the per-VM GPA into perf_test_args instead of storing it as a separate global variable. It's not obvious that guest_test_phys_mem holds a GPA, nor that it's connected/coupled with per_vcpu->gpa. No functional change intended. Signed-off-by: Sean Christopherson Reviewed-by: Ben Gardon Signed-off-by: David Matlack Message-Id: <20211111000310.1435032-9-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- .../testing/selftests/kvm/include/perf_test_util.h | 8 +------- tools/testing/selftests/kvm/lib/perf_test_util.c | 21 ++++++++------------- .../kvm/memslot_modification_stress_test.c | 2 +- 3 files changed, 10 insertions(+), 21 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index 20aec72fe7b8..d7cde1ab2a85 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -29,6 +29,7 @@ struct perf_test_vcpu_args { struct perf_test_args { struct kvm_vm *vm; uint64_t host_page_size; + uint64_t gpa; uint64_t guest_page_size; int wr_fract; @@ -37,13 +38,6 @@ struct perf_test_args { extern struct perf_test_args perf_test_args; -/* - * Guest physical memory offset of the testing memory slot. - * This will be set to the topmost valid physical address minus - * the test memory size. - */ -extern uint64_t guest_test_phys_mem; - struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, uint64_t vcpu_memory_bytes, int slots, enum vm_mem_backing_src_type backing_src); diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index d9c6bcb7964d..0fc2d834c1c7 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -10,8 +10,6 @@ struct perf_test_args perf_test_args; -uint64_t guest_test_phys_mem; - /* * Guest virtual memory offset of the testing memory slot. * Must not conflict with identity mapped test code. @@ -97,20 +95,18 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, guest_num_pages, vm_get_max_gfn(vm), vcpus, vcpu_memory_bytes); - guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * - pta->guest_page_size; - guest_test_phys_mem = align_down(guest_test_phys_mem, backing_src_pagesz); + pta->gpa = (vm_get_max_gfn(vm) - guest_num_pages) * pta->guest_page_size; + pta->gpa = align_down(pta->gpa, backing_src_pagesz); #ifdef __s390x__ /* Align to 1M (segment size) */ - guest_test_phys_mem = align_down(guest_test_phys_mem, 1 << 20); + pta->gpa = align_down(pta->gpa, 1 << 20); #endif - pr_info("guest physical test memory offset: 0x%lx\n", guest_test_phys_mem); + pr_info("guest physical test memory offset: 0x%lx\n", pta->gpa); /* Add extra memory slots for testing */ for (i = 0; i < slots; i++) { uint64_t region_pages = guest_num_pages / slots; - vm_paddr_t region_start = guest_test_phys_mem + - region_pages * pta->guest_page_size * i; + vm_paddr_t region_start = pta->gpa + region_pages * pta->guest_page_size * i; vm_userspace_mem_region_add(vm, backing_src, region_start, PERF_TEST_MEM_SLOT_INDEX + i, @@ -118,7 +114,7 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, } /* Do mapping for the demand paging memory slot */ - virt_map(vm, guest_test_virt_mem, guest_test_phys_mem, guest_num_pages); + virt_map(vm, guest_test_virt_mem, pta->gpa, guest_num_pages); ucall_init(vm, NULL); @@ -148,13 +144,12 @@ void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, (vcpu_id * vcpu_memory_bytes); vcpu_args->pages = vcpu_memory_bytes / pta->guest_page_size; - vcpu_args->gpa = guest_test_phys_mem + - (vcpu_id * vcpu_memory_bytes); + vcpu_args->gpa = pta->gpa + (vcpu_id * vcpu_memory_bytes); } else { vcpu_args->gva = guest_test_virt_mem; vcpu_args->pages = (vcpus * vcpu_memory_bytes) / pta->guest_page_size; - vcpu_args->gpa = guest_test_phys_mem; + vcpu_args->gpa = pta->gpa; } vcpu_args_set(vm, vcpu_id, 1, vcpu_id); diff --git a/tools/testing/selftests/kvm/memslot_modification_stress_test.c b/tools/testing/selftests/kvm/memslot_modification_stress_test.c index 4cfcafea9f5a..d105180d5e8c 100644 --- a/tools/testing/selftests/kvm/memslot_modification_stress_test.c +++ b/tools/testing/selftests/kvm/memslot_modification_stress_test.c @@ -80,7 +80,7 @@ static void add_remove_memslot(struct kvm_vm *vm, useconds_t delay, * Add the dummy memslot just below the perf_test_util memslot, which is * at the top of the guest physical address space. */ - gpa = guest_test_phys_mem - pages * vm_get_page_size(vm); + gpa = perf_test_args.gpa - pages * vm_get_page_size(vm); for (i = 0; i < nr_modifications; i++) { usleep(delay); -- cgit v1.2.3-70-g09d2 From a5ac0fd1b90ae811ba51da6a9928633bddefb792 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:07 +0000 Subject: KVM: selftests: Remove perf_test_args.host_page_size Remove perf_test_args.host_page_size and instead use getpagesize() so that it's somewhat obvious that, for tests that care about the host page size, they care about the system page size, not the hardware page size, e.g. that the logic is unchanged if hugepages are in play. No functional change intended. Signed-off-by: Sean Christopherson Reviewed-by: Ben Gardon Signed-off-by: David Matlack Message-Id: <20211111000310.1435032-10-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/include/perf_test_util.h | 1 - tools/testing/selftests/kvm/lib/perf_test_util.c | 3 +-- 2 files changed, 1 insertion(+), 3 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index d7cde1ab2a85..9348580dc5be 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -28,7 +28,6 @@ struct perf_test_vcpu_args { struct perf_test_args { struct kvm_vm *vm; - uint64_t host_page_size; uint64_t gpa; uint64_t guest_page_size; int wr_fract; diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index 0fc2d834c1c7..a0aded8cfce3 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -60,7 +60,6 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, pr_info("Testing guest mode: %s\n", vm_guest_mode_string(mode)); - pta->host_page_size = getpagesize(); /* * Snapshot the non-huge page size. This is used by the guest code to * access/dirty pages at the logging granularity. @@ -70,7 +69,7 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, guest_num_pages = vm_adjust_num_guest_pages(mode, (vcpus * vcpu_memory_bytes) / pta->guest_page_size); - TEST_ASSERT(vcpu_memory_bytes % pta->host_page_size == 0, + TEST_ASSERT(vcpu_memory_bytes % getpagesize() == 0, "Guest memory size is not host page size aligned."); TEST_ASSERT(vcpu_memory_bytes % pta->guest_page_size == 0, "Guest memory size is not guest page size aligned."); -- cgit v1.2.3-70-g09d2 From f5e8fe2a92e4923b63d1edd6ed53d9856b6515ce Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:08 +0000 Subject: KVM: selftests: Create VM with adjusted number of guest pages for perf tests Use the already computed guest_num_pages when creating the so called extra VM pages for a perf test, and add a comment explaining why the pages are allocated as extra pages. Signed-off-by: Sean Christopherson Reviewed-by: Ben Gardon Signed-off-by: David Matlack Message-Id: <20211111000310.1435032-11-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/lib/perf_test_util.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index a0aded8cfce3..b3154b5b0cfd 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -77,9 +77,13 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, "Guest memory cannot be evenly divided into %d slots.", slots); + /* + * Pass guest_num_pages to populate the page tables for test memory. + * The memory is also added to memslot 0, but that's a benign side + * effect as KVM allows aliasing HVAs in meslots. + */ vm = vm_create_with_vcpus(mode, vcpus, DEFAULT_GUEST_PHY_PAGES, - (vcpus * vcpu_memory_bytes) / pta->guest_page_size, - 0, guest_code, NULL); + guest_num_pages, 0, guest_code, NULL); pta->vm = vm; -- cgit v1.2.3-70-g09d2 From cf1d59300ab27af6a2e96b4882fe3d9a72b32b15 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:09 +0000 Subject: KVM: selftests: Fill per-vCPU struct during "perf_test" VM creation Fill the per-vCPU args when creating the perf_test VM instead of having the caller do so. This helps ensure that any adjustments to the number of pages (and thus vcpu_memory_bytes) are reflected in the per-VM args. Automatically filling the per-vCPU args will also allow a future patch to do the sync to the guest during creation. Signed-off-by: Sean Christopherson [Updated access_tracking_perf_test as well.] Signed-off-by: David Matlack Reviewed-by: Ben Gardon Message-Id: <20211111000310.1435032-12-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- .../selftests/kvm/access_tracking_perf_test.c | 5 +- tools/testing/selftests/kvm/demand_paging_test.c | 5 +- tools/testing/selftests/kvm/dirty_log_perf_test.c | 6 +- .../testing/selftests/kvm/include/perf_test_util.h | 6 +- tools/testing/selftests/kvm/lib/perf_test_util.c | 71 +++++++++++----------- .../kvm/memslot_modification_stress_test.c | 6 +- 6 files changed, 45 insertions(+), 54 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/access_tracking_perf_test.c b/tools/testing/selftests/kvm/access_tracking_perf_test.c index 5d95113c7b7c..fdef6c906388 100644 --- a/tools/testing/selftests/kvm/access_tracking_perf_test.c +++ b/tools/testing/selftests/kvm/access_tracking_perf_test.c @@ -332,10 +332,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) int vcpus = params->vcpus; vm = perf_test_create_vm(mode, vcpus, params->vcpu_memory_bytes, 1, - params->backing_src); - - perf_test_setup_vcpus(vm, vcpus, params->vcpu_memory_bytes, - !overlap_memory_access); + params->backing_src, !overlap_memory_access); vcpu_threads = create_vcpu_threads(vcpus); diff --git a/tools/testing/selftests/kvm/demand_paging_test.c b/tools/testing/selftests/kvm/demand_paging_test.c index 3c729a0a1ab1..0fee44f5e5ae 100644 --- a/tools/testing/selftests/kvm/demand_paging_test.c +++ b/tools/testing/selftests/kvm/demand_paging_test.c @@ -293,7 +293,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) int r; vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size, 1, - p->src_type); + p->src_type, p->partition_vcpu_memory_access); perf_test_args.wr_fract = 1; @@ -307,9 +307,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - perf_test_setup_vcpus(vm, nr_vcpus, guest_percpu_mem_size, - p->partition_vcpu_memory_access); - if (p->uffd_mode) { uffd_handler_threads = malloc(nr_vcpus * sizeof(*uffd_handler_threads)); diff --git a/tools/testing/selftests/kvm/dirty_log_perf_test.c b/tools/testing/selftests/kvm/dirty_log_perf_test.c index 7ffab5bd5ce5..62f9cc2a3146 100644 --- a/tools/testing/selftests/kvm/dirty_log_perf_test.c +++ b/tools/testing/selftests/kvm/dirty_log_perf_test.c @@ -186,7 +186,8 @@ static void run_test(enum vm_guest_mode mode, void *arg) struct timespec clear_dirty_log_total = (struct timespec){0}; vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size, - p->slots, p->backing_src); + p->slots, p->backing_src, + p->partition_vcpu_memory_access); perf_test_args.wr_fract = p->wr_fract; @@ -206,9 +207,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - perf_test_setup_vcpus(vm, nr_vcpus, guest_percpu_mem_size, - p->partition_vcpu_memory_access); - sync_global_to_guest(vm, perf_test_args); /* Start the iterations */ diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index 9348580dc5be..91804be1cf53 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -39,10 +39,8 @@ extern struct perf_test_args perf_test_args; struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, uint64_t vcpu_memory_bytes, int slots, - enum vm_mem_backing_src_type backing_src); + enum vm_mem_backing_src_type backing_src, + bool partition_vcpu_memory_access); void perf_test_destroy_vm(struct kvm_vm *vm); -void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, - uint64_t vcpu_memory_bytes, - bool partition_vcpu_memory_access); #endif /* SELFTEST_KVM_PERF_TEST_UTIL_H */ diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index b3154b5b0cfd..13c8bc22f4e1 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -48,9 +48,43 @@ static void guest_code(uint32_t vcpu_id) } } +void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, + uint64_t vcpu_memory_bytes, + bool partition_vcpu_memory_access) +{ + struct perf_test_args *pta = &perf_test_args; + struct perf_test_vcpu_args *vcpu_args; + int vcpu_id; + + for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { + vcpu_args = &pta->vcpu_args[vcpu_id]; + + vcpu_args->vcpu_id = vcpu_id; + if (partition_vcpu_memory_access) { + vcpu_args->gva = guest_test_virt_mem + + (vcpu_id * vcpu_memory_bytes); + vcpu_args->pages = vcpu_memory_bytes / + pta->guest_page_size; + vcpu_args->gpa = pta->gpa + (vcpu_id * vcpu_memory_bytes); + } else { + vcpu_args->gva = guest_test_virt_mem; + vcpu_args->pages = (vcpus * vcpu_memory_bytes) / + pta->guest_page_size; + vcpu_args->gpa = pta->gpa; + } + + vcpu_args_set(vm, vcpu_id, 1, vcpu_id); + + pr_debug("Added VCPU %d with test mem gpa [%lx, %lx)\n", + vcpu_id, vcpu_args->gpa, vcpu_args->gpa + + (vcpu_args->pages * pta->guest_page_size)); + } +} + struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, uint64_t vcpu_memory_bytes, int slots, - enum vm_mem_backing_src_type backing_src) + enum vm_mem_backing_src_type backing_src, + bool partition_vcpu_memory_access) { struct perf_test_args *pta = &perf_test_args; struct kvm_vm *vm; @@ -119,6 +153,8 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, /* Do mapping for the demand paging memory slot */ virt_map(vm, guest_test_virt_mem, pta->gpa, guest_num_pages); + perf_test_setup_vcpus(vm, vcpus, vcpu_memory_bytes, partition_vcpu_memory_access); + ucall_init(vm, NULL); return vm; @@ -129,36 +165,3 @@ void perf_test_destroy_vm(struct kvm_vm *vm) ucall_uninit(vm); kvm_vm_free(vm); } - -void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, - uint64_t vcpu_memory_bytes, - bool partition_vcpu_memory_access) -{ - struct perf_test_args *pta = &perf_test_args; - struct perf_test_vcpu_args *vcpu_args; - int vcpu_id; - - for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { - vcpu_args = &pta->vcpu_args[vcpu_id]; - - vcpu_args->vcpu_id = vcpu_id; - if (partition_vcpu_memory_access) { - vcpu_args->gva = guest_test_virt_mem + - (vcpu_id * vcpu_memory_bytes); - vcpu_args->pages = vcpu_memory_bytes / - pta->guest_page_size; - vcpu_args->gpa = pta->gpa + (vcpu_id * vcpu_memory_bytes); - } else { - vcpu_args->gva = guest_test_virt_mem; - vcpu_args->pages = (vcpus * vcpu_memory_bytes) / - pta->guest_page_size; - vcpu_args->gpa = pta->gpa; - } - - vcpu_args_set(vm, vcpu_id, 1, vcpu_id); - - pr_debug("Added VCPU %d with test mem gpa [%lx, %lx)\n", - vcpu_id, vcpu_args->gpa, vcpu_args->gpa + - (vcpu_args->pages * pta->guest_page_size)); - } -} diff --git a/tools/testing/selftests/kvm/memslot_modification_stress_test.c b/tools/testing/selftests/kvm/memslot_modification_stress_test.c index d105180d5e8c..27af0bb8deb7 100644 --- a/tools/testing/selftests/kvm/memslot_modification_stress_test.c +++ b/tools/testing/selftests/kvm/memslot_modification_stress_test.c @@ -105,16 +105,14 @@ static void run_test(enum vm_guest_mode mode, void *arg) int vcpu_id; vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size, 1, - VM_MEM_SRC_ANONYMOUS); + VM_MEM_SRC_ANONYMOUS, + p->partition_vcpu_memory_access); perf_test_args.wr_fract = 1; vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - perf_test_setup_vcpus(vm, nr_vcpus, guest_percpu_mem_size, - p->partition_vcpu_memory_access); - /* Export the shared variables to the guest */ sync_global_to_guest(vm, perf_test_args); -- cgit v1.2.3-70-g09d2 From 13bbc70329c8df003e64c4fbea8678f9db0e75d5 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Thu, 11 Nov 2021 00:03:10 +0000 Subject: KVM: selftests: Sync perf_test_args to guest during VM creation Copy perf_test_args to the guest during VM creation instead of relying on the caller to do so at their leisure. Ideally, tests wouldn't even be able to modify perf_test_args, i.e. they would have no motivation to do the sync, but enforcing that is arguably a net negative for readability. No functional change intended. [Set wr_fract=1 by default and add helper to override it since the new access_tracking_perf_test needs to set it dynamically.] Signed-off-by: Sean Christopherson Signed-off-by: David Matlack Reviewed-by: Ben Gardon Message-Id: <20211111000310.1435032-13-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/access_tracking_perf_test.c | 3 +-- tools/testing/selftests/kvm/demand_paging_test.c | 5 ----- tools/testing/selftests/kvm/dirty_log_perf_test.c | 4 +--- tools/testing/selftests/kvm/include/perf_test_util.h | 2 ++ tools/testing/selftests/kvm/lib/perf_test_util.c | 12 ++++++++++++ .../testing/selftests/kvm/memslot_modification_stress_test.c | 5 ----- 6 files changed, 16 insertions(+), 15 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/access_tracking_perf_test.c b/tools/testing/selftests/kvm/access_tracking_perf_test.c index fdef6c906388..5364a2ed7c68 100644 --- a/tools/testing/selftests/kvm/access_tracking_perf_test.c +++ b/tools/testing/selftests/kvm/access_tracking_perf_test.c @@ -277,8 +277,7 @@ static void run_iteration(struct kvm_vm *vm, int vcpus, const char *description) static void access_memory(struct kvm_vm *vm, int vcpus, enum access_type access, const char *description) { - perf_test_args.wr_fract = (access == ACCESS_READ) ? INT_MAX : 1; - sync_global_to_guest(vm, perf_test_args); + perf_test_set_wr_fract(vm, (access == ACCESS_READ) ? INT_MAX : 1); iteration_work = ITERATION_ACCESS_MEMORY; run_iteration(vm, vcpus, description); } diff --git a/tools/testing/selftests/kvm/demand_paging_test.c b/tools/testing/selftests/kvm/demand_paging_test.c index 0fee44f5e5ae..26f8fd8a57ec 100644 --- a/tools/testing/selftests/kvm/demand_paging_test.c +++ b/tools/testing/selftests/kvm/demand_paging_test.c @@ -295,8 +295,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size, 1, p->src_type, p->partition_vcpu_memory_access); - perf_test_args.wr_fract = 1; - demand_paging_size = get_backing_src_pagesz(p->src_type); guest_data_prototype = malloc(demand_paging_size); @@ -345,9 +343,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) } } - /* Export the shared variables to the guest */ - sync_global_to_guest(vm, perf_test_args); - pr_info("Finished creating vCPUs and starting uffd threads\n"); clock_gettime(CLOCK_MONOTONIC, &start); diff --git a/tools/testing/selftests/kvm/dirty_log_perf_test.c b/tools/testing/selftests/kvm/dirty_log_perf_test.c index 62f9cc2a3146..583b4d95aa98 100644 --- a/tools/testing/selftests/kvm/dirty_log_perf_test.c +++ b/tools/testing/selftests/kvm/dirty_log_perf_test.c @@ -189,7 +189,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) p->slots, p->backing_src, p->partition_vcpu_memory_access); - perf_test_args.wr_fract = p->wr_fract; + perf_test_set_wr_fract(vm, p->wr_fract); guest_num_pages = (nr_vcpus * guest_percpu_mem_size) >> vm_get_page_shift(vm); guest_num_pages = vm_adjust_num_guest_pages(mode, guest_num_pages); @@ -207,8 +207,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - sync_global_to_guest(vm, perf_test_args); - /* Start the iterations */ iteration = 0; host_quit = false; diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index 91804be1cf53..74e3622b3a6e 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -43,4 +43,6 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, bool partition_vcpu_memory_access); void perf_test_destroy_vm(struct kvm_vm *vm); +void perf_test_set_wr_fract(struct kvm_vm *vm, int wr_fract); + #endif /* SELFTEST_KVM_PERF_TEST_UTIL_H */ diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index 13c8bc22f4e1..77f9eb5667c9 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -94,6 +94,9 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, pr_info("Testing guest mode: %s\n", vm_guest_mode_string(mode)); + /* By default vCPUs will write to memory. */ + pta->wr_fract = 1; + /* * Snapshot the non-huge page size. This is used by the guest code to * access/dirty pages at the logging granularity. @@ -157,6 +160,9 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, ucall_init(vm, NULL); + /* Export the shared variables to the guest. */ + sync_global_to_guest(vm, perf_test_args); + return vm; } @@ -165,3 +171,9 @@ void perf_test_destroy_vm(struct kvm_vm *vm) ucall_uninit(vm); kvm_vm_free(vm); } + +void perf_test_set_wr_fract(struct kvm_vm *vm, int wr_fract) +{ + perf_test_args.wr_fract = wr_fract; + sync_global_to_guest(vm, perf_test_args); +} diff --git a/tools/testing/selftests/kvm/memslot_modification_stress_test.c b/tools/testing/selftests/kvm/memslot_modification_stress_test.c index 27af0bb8deb7..df431d0da1ee 100644 --- a/tools/testing/selftests/kvm/memslot_modification_stress_test.c +++ b/tools/testing/selftests/kvm/memslot_modification_stress_test.c @@ -108,14 +108,9 @@ static void run_test(enum vm_guest_mode mode, void *arg) VM_MEM_SRC_ANONYMOUS, p->partition_vcpu_memory_access); - perf_test_args.wr_fract = 1; - vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - /* Export the shared variables to the guest */ - sync_global_to_guest(vm, perf_test_args); - pr_info("Finished creating vCPUs\n"); for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) -- cgit v1.2.3-70-g09d2 From 81bcb26172a8f00840e0ca44277272dcb673887a Mon Sep 17 00:00:00 2001 From: David Matlack Date: Thu, 11 Nov 2021 00:12:55 +0000 Subject: KVM: selftests: Move vCPU thread creation and joining to common helpers Move vCPU thread creation and joining to common helper functions. This is in preparation for the next commit which ensures that all vCPU threads are fully created before entering guest mode on any one vCPU. No functional change intended. Signed-off-by: David Matlack Reviewed-by: Ben Gardon Message-Id: <20211111001257.1446428-3-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- .../selftests/kvm/access_tracking_perf_test.c | 40 +++---------------- tools/testing/selftests/kvm/demand_paging_test.c | 25 ++---------- tools/testing/selftests/kvm/dirty_log_perf_test.c | 19 ++------- .../testing/selftests/kvm/include/perf_test_util.h | 5 +++ tools/testing/selftests/kvm/lib/perf_test_util.c | 46 ++++++++++++++++++++++ .../kvm/memslot_modification_stress_test.c | 22 ++--------- 6 files changed, 67 insertions(+), 90 deletions(-) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/access_tracking_perf_test.c b/tools/testing/selftests/kvm/access_tracking_perf_test.c index 7f25a06e19c9..d8909032317a 100644 --- a/tools/testing/selftests/kvm/access_tracking_perf_test.c +++ b/tools/testing/selftests/kvm/access_tracking_perf_test.c @@ -215,9 +215,8 @@ static bool spin_wait_for_next_iteration(int *current_iteration) return true; } -static void *vcpu_thread_main(void *arg) +static void vcpu_thread_main(struct perf_test_vcpu_args *vcpu_args) { - struct perf_test_vcpu_args *vcpu_args = arg; struct kvm_vm *vm = perf_test_args.vm; int vcpu_id = vcpu_args->vcpu_id; int current_iteration = 0; @@ -235,8 +234,6 @@ static void *vcpu_thread_main(void *arg) vcpu_last_completed_iteration[vcpu_id] = current_iteration; } - - return NULL; } static void spin_wait_for_vcpu(int vcpu_id, int target_iteration) @@ -295,43 +292,16 @@ static void mark_memory_idle(struct kvm_vm *vm, int vcpus) run_iteration(vm, vcpus, "Mark memory idle"); } -static pthread_t *create_vcpu_threads(int vcpus) -{ - pthread_t *vcpu_threads; - int i; - - vcpu_threads = malloc(vcpus * sizeof(vcpu_threads[0])); - TEST_ASSERT(vcpu_threads, "Failed to allocate vcpu_threads."); - - for (i = 0; i < vcpus; i++) - pthread_create(&vcpu_threads[i], NULL, vcpu_thread_main, - &perf_test_args.vcpu_args[i]); - - return vcpu_threads; -} - -static void terminate_vcpu_threads(pthread_t *vcpu_threads, int vcpus) -{ - int i; - - /* Set done to signal the vCPU threads to exit */ - done = true; - - for (i = 0; i < vcpus; i++) - pthread_join(vcpu_threads[i], NULL); -} - static void run_test(enum vm_guest_mode mode, void *arg) { struct test_params *params = arg; struct kvm_vm *vm; - pthread_t *vcpu_threads; int vcpus = params->vcpus; vm = perf_test_create_vm(mode, vcpus, params->vcpu_memory_bytes, 1, params->backing_src, !overlap_memory_access); - vcpu_threads = create_vcpu_threads(vcpus); + perf_test_start_vcpu_threads(vcpus, vcpu_thread_main); pr_info("\n"); access_memory(vm, vcpus, ACCESS_WRITE, "Populating memory"); @@ -346,8 +316,10 @@ static void run_test(enum vm_guest_mode mode, void *arg) mark_memory_idle(vm, vcpus); access_memory(vm, vcpus, ACCESS_READ, "Reading from idle memory"); - terminate_vcpu_threads(vcpu_threads, vcpus); - free(vcpu_threads); + /* Set done to signal the vCPU threads to exit */ + done = true; + + perf_test_join_vcpu_threads(vcpus); perf_test_destroy_vm(vm); } diff --git a/tools/testing/selftests/kvm/demand_paging_test.c b/tools/testing/selftests/kvm/demand_paging_test.c index 26f8fd8a57ec..6a719d065599 100644 --- a/tools/testing/selftests/kvm/demand_paging_test.c +++ b/tools/testing/selftests/kvm/demand_paging_test.c @@ -42,10 +42,9 @@ static uint64_t guest_percpu_mem_size = DEFAULT_PER_VCPU_MEM_SIZE; static size_t demand_paging_size; static char *guest_data_prototype; -static void *vcpu_worker(void *data) +static void vcpu_worker(struct perf_test_vcpu_args *vcpu_args) { int ret; - struct perf_test_vcpu_args *vcpu_args = (struct perf_test_vcpu_args *)data; int vcpu_id = vcpu_args->vcpu_id; struct kvm_vm *vm = perf_test_args.vm; struct kvm_run *run; @@ -68,8 +67,6 @@ static void *vcpu_worker(void *data) ts_diff = timespec_elapsed(start); PER_VCPU_DEBUG("vCPU %d execution time: %ld.%.9lds\n", vcpu_id, ts_diff.tv_sec, ts_diff.tv_nsec); - - return NULL; } static int handle_uffd_page_request(int uffd_mode, int uffd, uint64_t addr) @@ -282,7 +279,6 @@ struct test_params { static void run_test(enum vm_guest_mode mode, void *arg) { struct test_params *p = arg; - pthread_t *vcpu_threads; pthread_t *uffd_handler_threads = NULL; struct uffd_handler_args *uffd_args = NULL; struct timespec start; @@ -302,9 +298,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) "Failed to allocate buffer for guest data pattern"); memset(guest_data_prototype, 0xAB, demand_paging_size); - vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); - TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - if (p->uffd_mode) { uffd_handler_threads = malloc(nr_vcpus * sizeof(*uffd_handler_threads)); @@ -346,22 +339,11 @@ static void run_test(enum vm_guest_mode mode, void *arg) pr_info("Finished creating vCPUs and starting uffd threads\n"); clock_gettime(CLOCK_MONOTONIC, &start); - - for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) { - pthread_create(&vcpu_threads[vcpu_id], NULL, vcpu_worker, - &perf_test_args.vcpu_args[vcpu_id]); - } - + perf_test_start_vcpu_threads(nr_vcpus, vcpu_worker); pr_info("Started all vCPUs\n"); - /* Wait for the vcpu threads to quit */ - for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) { - pthread_join(vcpu_threads[vcpu_id], NULL); - PER_VCPU_DEBUG("Joined thread for vCPU %d\n", vcpu_id); - } - + perf_test_join_vcpu_threads(nr_vcpus); ts_diff = timespec_elapsed(start); - pr_info("All vCPU threads joined\n"); if (p->uffd_mode) { @@ -385,7 +367,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) perf_test_destroy_vm(vm); free(guest_data_prototype); - free(vcpu_threads); if (p->uffd_mode) { free(uffd_handler_threads); free(uffd_args); diff --git a/tools/testing/selftests/kvm/dirty_log_perf_test.c b/tools/testing/selftests/kvm/dirty_log_perf_test.c index 583b4d95aa98..1954b964d1cf 100644 --- a/tools/testing/selftests/kvm/dirty_log_perf_test.c +++ b/tools/testing/selftests/kvm/dirty_log_perf_test.c @@ -31,7 +31,7 @@ static bool host_quit; static int iteration; static int vcpu_last_completed_iteration[KVM_MAX_VCPUS]; -static void *vcpu_worker(void *data) +static void vcpu_worker(struct perf_test_vcpu_args *vcpu_args) { int ret; struct kvm_vm *vm = perf_test_args.vm; @@ -41,7 +41,6 @@ static void *vcpu_worker(void *data) struct timespec ts_diff; struct timespec total = (struct timespec){0}; struct timespec avg; - struct perf_test_vcpu_args *vcpu_args = (struct perf_test_vcpu_args *)data; int vcpu_id = vcpu_args->vcpu_id; run = vcpu_state(vm, vcpu_id); @@ -83,8 +82,6 @@ static void *vcpu_worker(void *data) pr_debug("\nvCPU %d dirtied 0x%lx pages over %d iterations in %ld.%.9lds. (Avg %ld.%.9lds/iteration)\n", vcpu_id, pages_count, vcpu_last_completed_iteration[vcpu_id], total.tv_sec, total.tv_nsec, avg.tv_sec, avg.tv_nsec); - - return NULL; } struct test_params { @@ -170,7 +167,6 @@ static void free_bitmaps(unsigned long *bitmaps[], int slots) static void run_test(enum vm_guest_mode mode, void *arg) { struct test_params *p = arg; - pthread_t *vcpu_threads; struct kvm_vm *vm; unsigned long **bitmaps; uint64_t guest_num_pages; @@ -204,20 +200,15 @@ static void run_test(enum vm_guest_mode mode, void *arg) vm_enable_cap(vm, &cap); } - vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); - TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - /* Start the iterations */ iteration = 0; host_quit = false; clock_gettime(CLOCK_MONOTONIC, &start); - for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) { + for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) vcpu_last_completed_iteration[vcpu_id] = -1; - pthread_create(&vcpu_threads[vcpu_id], NULL, vcpu_worker, - &perf_test_args.vcpu_args[vcpu_id]); - } + perf_test_start_vcpu_threads(nr_vcpus, vcpu_worker); /* Allow the vCPUs to populate memory */ pr_debug("Starting iteration %d - Populating\n", iteration); @@ -286,8 +277,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) /* Tell the vcpu thread to quit */ host_quit = true; - for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) - pthread_join(vcpu_threads[vcpu_id], NULL); + perf_test_join_vcpu_threads(nr_vcpus); avg = timespec_div(get_dirty_log_total, p->iterations); pr_info("Get dirty log over %lu iterations took %ld.%.9lds. (Avg %ld.%.9lds/iteration)\n", @@ -302,7 +292,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) } free_bitmaps(bitmaps, p->slots); - free(vcpu_threads); perf_test_destroy_vm(vm); } diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index 74e3622b3a6e..a86f953d8d36 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -8,6 +8,8 @@ #ifndef SELFTEST_KVM_PERF_TEST_UTIL_H #define SELFTEST_KVM_PERF_TEST_UTIL_H +#include + #include "kvm_util.h" /* Default guest test virtual memory offset */ @@ -45,4 +47,7 @@ void perf_test_destroy_vm(struct kvm_vm *vm); void perf_test_set_wr_fract(struct kvm_vm *vm, int wr_fract); +void perf_test_start_vcpu_threads(int vcpus, void (*vcpu_fn)(struct perf_test_vcpu_args *)); +void perf_test_join_vcpu_threads(int vcpus); + #endif /* SELFTEST_KVM_PERF_TEST_UTIL_H */ diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index 77f9eb5667c9..d646477ed16a 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -16,6 +16,20 @@ struct perf_test_args perf_test_args; */ static uint64_t guest_test_virt_mem = DEFAULT_GUEST_TEST_MEM; +struct vcpu_thread { + /* The id of the vCPU. */ + int vcpu_id; + + /* The pthread backing the vCPU. */ + pthread_t thread; +}; + +/* The vCPU threads involved in this test. */ +static struct vcpu_thread vcpu_threads[KVM_MAX_VCPUS]; + +/* The function run by each vCPU thread, as provided by the test. */ +static void (*vcpu_thread_fn)(struct perf_test_vcpu_args *); + /* * Continuously write to the first 8 bytes of each page in the * specified region. @@ -177,3 +191,35 @@ void perf_test_set_wr_fract(struct kvm_vm *vm, int wr_fract) perf_test_args.wr_fract = wr_fract; sync_global_to_guest(vm, perf_test_args); } + +static void *vcpu_thread_main(void *data) +{ + struct vcpu_thread *vcpu = data; + + vcpu_thread_fn(&perf_test_args.vcpu_args[vcpu->vcpu_id]); + + return NULL; +} + +void perf_test_start_vcpu_threads(int vcpus, void (*vcpu_fn)(struct perf_test_vcpu_args *)) +{ + int vcpu_id; + + vcpu_thread_fn = vcpu_fn; + + for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { + struct vcpu_thread *vcpu = &vcpu_threads[vcpu_id]; + + vcpu->vcpu_id = vcpu_id; + + pthread_create(&vcpu->thread, NULL, vcpu_thread_main, vcpu); + } +} + +void perf_test_join_vcpu_threads(int vcpus) +{ + int vcpu_id; + + for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) + pthread_join(vcpu_threads[vcpu_id].thread, NULL); +} diff --git a/tools/testing/selftests/kvm/memslot_modification_stress_test.c b/tools/testing/selftests/kvm/memslot_modification_stress_test.c index df431d0da1ee..5bd0b076f57f 100644 --- a/tools/testing/selftests/kvm/memslot_modification_stress_test.c +++ b/tools/testing/selftests/kvm/memslot_modification_stress_test.c @@ -36,11 +36,9 @@ static uint64_t guest_percpu_mem_size = DEFAULT_PER_VCPU_MEM_SIZE; static bool run_vcpus = true; -static void *vcpu_worker(void *data) +static void vcpu_worker(struct perf_test_vcpu_args *vcpu_args) { int ret; - struct perf_test_vcpu_args *vcpu_args = - (struct perf_test_vcpu_args *)data; int vcpu_id = vcpu_args->vcpu_id; struct kvm_vm *vm = perf_test_args.vm; struct kvm_run *run; @@ -59,8 +57,6 @@ static void *vcpu_worker(void *data) "Invalid guest sync status: exit_reason=%s\n", exit_reason_str(run->exit_reason)); } - - return NULL; } struct memslot_antagonist_args { @@ -100,22 +96,15 @@ struct test_params { static void run_test(enum vm_guest_mode mode, void *arg) { struct test_params *p = arg; - pthread_t *vcpu_threads; struct kvm_vm *vm; - int vcpu_id; vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size, 1, VM_MEM_SRC_ANONYMOUS, p->partition_vcpu_memory_access); - vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); - TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - pr_info("Finished creating vCPUs\n"); - for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) - pthread_create(&vcpu_threads[vcpu_id], NULL, vcpu_worker, - &perf_test_args.vcpu_args[vcpu_id]); + perf_test_start_vcpu_threads(nr_vcpus, vcpu_worker); pr_info("Started all vCPUs\n"); @@ -124,16 +113,11 @@ static void run_test(enum vm_guest_mode mode, void *arg) run_vcpus = false; - /* Wait for the vcpu threads to quit */ - for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) - pthread_join(vcpu_threads[vcpu_id], NULL); - + perf_test_join_vcpu_threads(nr_vcpus); pr_info("All vCPU threads joined\n"); ucall_uninit(vm); kvm_vm_free(vm); - - free(vcpu_threads); } static void help(char *name) -- cgit v1.2.3-70-g09d2 From 89d9a43c1d2d3d703fae25c990a1d98dd178dd17 Mon Sep 17 00:00:00 2001 From: David Matlack Date: Thu, 11 Nov 2021 00:12:56 +0000 Subject: KVM: selftests: Wait for all vCPU to be created before entering guest mode Thread creation requires taking the mmap_sem in write mode, which causes vCPU threads running in guest mode to block while they are populating memory. Fix this by waiting for all vCPU threads to be created and start running before entering guest mode on any one vCPU thread. This substantially improves the "Populate memory time" when using 1GiB pages since it allows all vCPUs to zero pages in parallel rather than blocking because a writer is waiting (which is waiting for another vCPU that is busy zeroing a 1GiB page). Before: $ ./dirty_log_perf_test -v256 -s anonymous_hugetlb_1gb ... Populate memory time: 52.811184013s After: $ ./dirty_log_perf_test -v256 -s anonymous_hugetlb_1gb ... Populate memory time: 10.204573342s Signed-off-by: David Matlack Message-Id: <20211111001257.1446428-4-dmatlack@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/lib/perf_test_util.c | 26 ++++++++++++++++++++++++ 1 file changed, 26 insertions(+) (limited to 'tools/testing/selftests/kvm/lib/perf_test_util.c') diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c index d646477ed16a..722df3a28791 100644 --- a/tools/testing/selftests/kvm/lib/perf_test_util.c +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -22,6 +22,9 @@ struct vcpu_thread { /* The pthread backing the vCPU. */ pthread_t thread; + + /* Set to true once the vCPU thread is up and running. */ + bool running; }; /* The vCPU threads involved in this test. */ @@ -30,6 +33,9 @@ static struct vcpu_thread vcpu_threads[KVM_MAX_VCPUS]; /* The function run by each vCPU thread, as provided by the test. */ static void (*vcpu_thread_fn)(struct perf_test_vcpu_args *); +/* Set to true once all vCPU threads are up and running. */ +static bool all_vcpu_threads_running; + /* * Continuously write to the first 8 bytes of each page in the * specified region. @@ -196,6 +202,17 @@ static void *vcpu_thread_main(void *data) { struct vcpu_thread *vcpu = data; + WRITE_ONCE(vcpu->running, true); + + /* + * Wait for all vCPU threads to be up and running before calling the test- + * provided vCPU thread function. This prevents thread creation (which + * requires taking the mmap_sem in write mode) from interfering with the + * guest faulting in its memory. + */ + while (!READ_ONCE(all_vcpu_threads_running)) + ; + vcpu_thread_fn(&perf_test_args.vcpu_args[vcpu->vcpu_id]); return NULL; @@ -206,14 +223,23 @@ void perf_test_start_vcpu_threads(int vcpus, void (*vcpu_fn)(struct perf_test_vc int vcpu_id; vcpu_thread_fn = vcpu_fn; + WRITE_ONCE(all_vcpu_threads_running, false); for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { struct vcpu_thread *vcpu = &vcpu_threads[vcpu_id]; vcpu->vcpu_id = vcpu_id; + WRITE_ONCE(vcpu->running, false); pthread_create(&vcpu->thread, NULL, vcpu_thread_main, vcpu); } + + for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { + while (!READ_ONCE(vcpu_threads[vcpu_id].running)) + ; + } + + WRITE_ONCE(all_vcpu_threads_running, true); } void perf_test_join_vcpu_threads(int vcpus) -- cgit v1.2.3-70-g09d2