Re: [PATCH v3 9/9] selftests: kvm: Add guest_memfd_preservation_test

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Stage 2 test failure due to re-creation of preserved KVM resources (memslots and vCPU)
--

commit 323482e513bb6a4836ceedaaec02c412bc10f974
Author: Tarun Sahu <tarunsahu@google.com>

selftests: kvm: Add guest_memfd_preservation_test

This commit adds a new KVM selftest to verify that guest memory backed
by guest_memfd is preserved properly across kexec via LUO. It introduces
a two-stage test to save state before kexec and verify it afterwards.

> diff --git a/tools/testing/selftests/kvm/guest_memfd_preservation_test.c b/tools/testing/selftests/kvm/guest_memfd_preservation_test.c
> new file mode 100644
> index 0000000000000..c0a20e788ba79
> --- /dev/null
> +++ b/tools/testing/selftests/kvm/guest_memfd_preservation_test.c

[ ... ]

> +static void run_stage_2(int luo_fd, int state_session_fd)
> +{
> +	int retrieved_vm_fd, retrieved_gmem_fd, session_fd, stage;
> +	struct vm_shape shape = VM_SHAPE_DEFAULT;
> +	const uint64_t gpa = SZ_4G;
> +	struct kvm_vcpu *vcpu;
> +	const int slot = 1;
> +	struct kvm_vm *vm;

[ ... ]

> +	retrieved_gmem_fd = luo_session_retrieve_fd(session_fd, GMEM_TOKEN);
> +	TEST_ASSERT(retrieved_gmem_fd >= 0, "Failed to retrieve guest_memfd file descriptor");
> +
> +	vm = vm_create_from_fd(retrieved_vm_fd, shape);
> +
> +	u64 nr_pages = 2048; /* 8MB is plenty for slot0 pages */
> +
> +	vm_userspace_mem_region_add(vm, VM_MEM_SRC_ANONYMOUS, 0, 0, nr_pages, 0);

[Severity: High]
Does this ioctl fail with -EINVAL because slot 0 already exists in the
preserved KVM VM with a different size? The preserved VM from stage 1 retains
its kernel-side state, which we attempt to blindly re-create here without
respecting the preserved state.

> +	kvm_vm_elf_load(vm, program_invocation_name);
> +
> +	for (int i = 0; i < NR_MEM_REGIONS; i++)
> +		vm->memslots[i] = 0;
> +
> +	struct userspace_mem_region *slot0 = memslot2region(vm, 0);
> +
> +	ucall_init(vm, slot0->region.guest_phys_addr + slot0->region.memory_size);
> +
> +	vm_set_user_memory_region2(vm, slot, KVM_MEM_GUEST_MEMFD, gpa, GMEM_SIZE, NULL,
> +				   retrieved_gmem_fd, 0);

[Severity: High]
Could this fail with -EINVAL because modifying an existing private
KVM_MEM_GUEST_MEMFD memslot is explicitly forbidden by KVM?

> +
> +	for (size_t i = 0; i < GMEM_SIZE; i += page_size)
> +		virt_pg_map(vm, gpa + i, gpa + i);
> +
> +	vcpu = vm_vcpu_add(vm, 0, guest_code_phase2);

[Severity: High]
Will this call fail with -EEXIST since vCPU ID 0 was already created
in Stage 1 and remains preserved within the retrieved KVM VM?

> +	kvm_arch_vm_finalize_vcpus(vm);
> +
> +	vcpu_args_set(vcpu, 3, gpa, GMEM_SIZE, DATA_SIZE);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260622184851.2309827-1-tarunsahu@google.com?part=9