| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Mon, 6 Apr 2020 14:10:53 -0300
From: Wainer dos Santos Moschetta <wainersm@...hat.com>
To: Andrew Jones <drjones@...hat.com>
Cc: pbonzini@...hat.com, kvm@...r.kernel.org, david@...hat.com,
linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org
Subject: Re: [PATCH v2 2/2] selftests: kvm: Add mem_slot_test test
On 4/4/20 4:32 AM, Andrew Jones wrote:
> On Fri, Apr 03, 2020 at 02:24:28PM -0300, Wainer dos Santos Moschetta wrote:
>> This patch introduces the mem_slot_test test which checks
>> an VM can have added memory slots up to the limit defined in
>> KVM_CAP_NR_MEMSLOTS. Then attempt to add one more slot to
>> verify it fails as expected.
>>
>> Signed-off-by: Wainer dos Santos Moschetta <wainersm@...hat.com>
>> ---
>> tools/testing/selftests/kvm/.gitignore | 1 +
>> tools/testing/selftests/kvm/Makefile | 3 +
>> tools/testing/selftests/kvm/mem_slot_test.c | 85 +++++++++++++++++++++
>> 3 files changed, 89 insertions(+)
>> create mode 100644 tools/testing/selftests/kvm/mem_slot_test.c
>>
>> diff --git a/tools/testing/selftests/kvm/.gitignore b/tools/testing/selftests/kvm/.gitignore
>> index 16877c3daabf..232f24d6931a 100644
>> --- a/tools/testing/selftests/kvm/.gitignore
>> +++ b/tools/testing/selftests/kvm/.gitignore
>> @@ -22,3 +22,4 @@
>> /dirty_log_test
>> /kvm_create_max_vcpus
>> /steal_time
>> +/mem_slot_test
>> diff --git a/tools/testing/selftests/kvm/Makefile b/tools/testing/selftests/kvm/Makefile
>> index 712a2ddd2a27..69b44178f48b 100644
>> --- a/tools/testing/selftests/kvm/Makefile
>> +++ b/tools/testing/selftests/kvm/Makefile
>> @@ -33,12 +33,14 @@ TEST_GEN_PROGS_x86_64 += demand_paging_test
>> TEST_GEN_PROGS_x86_64 += dirty_log_test
>> TEST_GEN_PROGS_x86_64 += kvm_create_max_vcpus
>> TEST_GEN_PROGS_x86_64 += steal_time
>> +TEST_GEN_PROGS_x86_64 += mem_slot_test
>>
>> TEST_GEN_PROGS_aarch64 += clear_dirty_log_test
>> TEST_GEN_PROGS_aarch64 += demand_paging_test
>> TEST_GEN_PROGS_aarch64 += dirty_log_test
>> TEST_GEN_PROGS_aarch64 += kvm_create_max_vcpus
>> TEST_GEN_PROGS_aarch64 += steal_time
>> +TEST_GEN_PROGS_aarch64 += mem_slot_test
>>
> kvm selftests has a bad case of OCD when it comes to lists of tests. In
> the .gitignore and the Makefile we keep our tests in alphabetical order.
> Maybe we should stop, because it's a bit annoying to maintain, but my
> personal OCD won't allow it to be on my watch. Please fix the above
> three lists.
I will fix it on v3.
Kind of related... has ever been discussed a naming convention for kvm
selftests? It would allow the use of regex on both .gitignore and
Makefile...and bye-bye those sorted lists.
>
>> TEST_GEN_PROGS_s390x = s390x/memop
>> TEST_GEN_PROGS_s390x += s390x/resets
>> @@ -46,6 +48,7 @@ TEST_GEN_PROGS_s390x += s390x/sync_regs_test
>> TEST_GEN_PROGS_s390x += demand_paging_test
>> TEST_GEN_PROGS_s390x += dirty_log_test
>> TEST_GEN_PROGS_s390x += kvm_create_max_vcpus
>> +TEST_GEN_PROGS_s390x += mem_slot_test
>>
>> TEST_GEN_PROGS += $(TEST_GEN_PROGS_$(UNAME_M))
>> LIBKVM += $(LIBKVM_$(UNAME_M))
>> diff --git a/tools/testing/selftests/kvm/mem_slot_test.c b/tools/testing/selftests/kvm/mem_slot_test.c
>> new file mode 100644
>> index 000000000000..eef6f506f41d
>> --- /dev/null
>> +++ b/tools/testing/selftests/kvm/mem_slot_test.c
>> @@ -0,0 +1,85 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * mem_slot_test
>> + *
>> + * Copyright (C) 2020, Red Hat, Inc.
>> + *
>> + * Test suite for memory region operations.
>> + */
>> +#define _GNU_SOURCE /* for program_invocation_short_name */
>> +#include <linux/kvm.h>
>> +#include <sys/mman.h>
>> +
>> +#include "test_util.h"
>> +#include "kvm_util.h"
>> +
>> +/*
>> + * Test it can be added memory slots up to KVM_CAP_NR_MEMSLOTS, then any
>> + * tentative to add further slots should fail.
>> + */
>> +static void test_add_max_slots(void)
>> +{
>> + struct kvm_vm *vm;
>> + uint32_t max_mem_slots;
>> + uint32_t slot;
>> + uint64_t mem_reg_npages;
>> + uint64_t mem_reg_size;
>> + uint32_t mem_reg_flags;
>> + uint64_t guest_addr;
>> + int ret;
>> +
>> + max_mem_slots = kvm_check_cap(KVM_CAP_NR_MEMSLOTS);
>> + TEST_ASSERT(max_mem_slots > 0,
>> + "KVM_CAP_NR_MEMSLOTS should be greater than 0");
>> + pr_info("Allowed number of memory slots: %i\n", max_mem_slots);
>> +
>> + vm = vm_create(VM_MODE_DEFAULT, 0, O_RDWR);
>> +
>> + /*
>> + * Uses 1MB sized/aligned memory region since this is the minimal
>> + * required on s390x.
>> + */
>> + mem_reg_size = 0x100000;
>> + mem_reg_npages = vm_calc_num_guest_pages(VM_MODE_DEFAULT, mem_reg_size);
>> +
>> + mem_reg_flags = kvm_check_cap(KVM_CAP_READONLY_MEM) ? KVM_MEM_READONLY :
>> + KVM_MEM_LOG_DIRTY_PAGES;
> I still don't see why we're setting a flag at all, and now we're setting
> different flags depending on what's available, so the test isn't the
> same for every environment. I would just have mem->flags = 0 for this
> test.
I thought I had to set a memory flag always. If mem->flags = 0 works
across the arches, then I change this on v3.
>
>> +
>> + guest_addr = 0x0;
>> +
>> + /* Check it can be added memory slots up to the maximum allowed */
>> + pr_info("Adding slots 0..%i, each memory region with %ldK size\n",
>> + (max_mem_slots - 1), mem_reg_size >> 10);
>> + for (slot = 0; slot < max_mem_slots; slot++) {
>> + vm_userspace_mem_region_add(vm, VM_MEM_SRC_ANONYMOUS,
>> + guest_addr, slot, mem_reg_npages,
>> + mem_reg_flags);
>> + guest_addr += mem_reg_size;
>> + }
>> +
>> + /* Check it cannot be added memory slots beyond the limit */
>> + void *mem = mmap(NULL, mem_reg_size, PROT_READ | PROT_WRITE,
>> + MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
>> + TEST_ASSERT(mem != NULL, "Failed to mmap() host");
> This should be testing mem != MAP_FAILED
Ok.
>
>> +
>> + struct kvm_userspace_memory_region kvm_region = {
>> + .slot = slot,
>> + .flags = mem_reg_flags,
>> + .guest_phys_addr = guest_addr,
>> + .memory_size = mem_reg_size,
>> + .userspace_addr = (uint64_t) mem,
>> + };
> Declaring kvm_region in the middle of the block. I don't really care
> myself, but it's inconsistent with all the other variables which are
> declared at the top.
Makes sense.
>
>> +
>> + ret = ioctl(vm_get_fd(vm), KVM_SET_USER_MEMORY_REGION, &kvm_region);
>> + TEST_ASSERT(ret == -1, "Adding one more memory slot should fail");
>> + TEST_ASSERT(errno == EINVAL, "Should return EINVAL errno");
> Please make the second assert message more specific. Or better would be
> to combine the asserts
>
> TEST_ASSERT(ret == -1 && errno == EINVAL, "Adding one more memory slot should fail with EINVAL");
Yeah, I was unsure about and'ing the checks. I will change it on v3.
Thanks!
Wainer
>
>> +
>> + munmap(mem, mem_reg_size);
>> + kvm_vm_free(vm);
>> +}
>> +
>> +int main(int argc, char *argv[])
>> +{
>> + test_add_max_slots();
>> + return 0;
>> +}
>> --
>> 2.17.2
>>
> Thanks,
> drew
Powered by blists - more mailing lists