Re: [PATCH v4 1/3] KVM: PPC: selftests: add support for powerpc

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] The patch relies on a missing KVM selftests core infrastructure patch, leading to undefined function calls and a build failure.
--

commit abd17de4492c780fa3b8540774bca368e3f9baa2
Author: Nicholas Piggin <npiggin@gmail.com>

KVM: PPC: selftests: add support for powerpc

This commit implements KVM selftests support for powerpc (Book3S-64),
including virtual memory for the radix MMU and guest interrupt handling
in real-mode. It also updates various types and adapts to the new page
table allocator APIs.

Link: https://lore.kernel.org/linuxppc-dev/cover.1781093720.git.ritesh.list@gmail.com/

> diff --git a/tools/testing/selftests/kvm/lib/powerpc/processor.c b/tools/testing/selftests/kvm/lib/powerpc/processor.c
> new file mode 100644
> index 0000000000000..57f8d90b55e3d
> --- /dev/null
> +++ b/tools/testing/selftests/kvm/lib/powerpc/processor.c

[ ... ]

> +void virt_arch_pgd_alloc(struct kvm_vm *vm)
> +{
> +	struct kvm_ppc_mmuv3_cfg mmu_cfg;
> +	gpa_t prtb, pgtb;
> +	size_t pgd_pages;
> +
> +	TEST_ASSERT((vm->mode == VM_MODE_P52V52_4K) ||
> +		    (vm->mode == VM_MODE_P52V52_64K),
> +		    "Unsupported guest mode, mode: 0x%x", vm->mode);
> +
> +	if (vm->mmu.pgd_created)
> +		return;
> +
> +	prtb = vm_alloc_page_table(vm);
> +	vm->arch.prtb = prtb;
> +
> +	pgd_pages = (1UL << (RADIX_PGD_INDEX_SIZE + 3)) >> vm->page_shift;
> +	if (!pgd_pages)
> +		pgd_pages = 1;
> +	pgtb = vm_alloc_page_table_pages(vm, pgd_pages);

[Severity: Low]
Are we missing a prerequisite patch for the selftests core infrastructure?
Building the tests for powerpc fails in virt_arch_pgd_alloc() because
vm_alloc_page_table_pages() is undefined.

> +	vm->mmu.pgd = pgtb;
> +
> +	/* Set the base page directory in the proc table */
> +	set_radix_proc_table(vm, 0, pgtb);

[ ... ]

> +static gpa_t __vm_alloc_pt(struct kvm_vm *vm, u64 pt_shift)
> +{
> +	gpa_t pt;
> +
> +	if (pt_shift >= vm->page_shift) {
> +		size_t pt_pages = 1ULL << (pt_shift - vm->page_shift);
> +
> +		pt = vm_alloc_page_table_pages(vm, pt_pages);

[Severity: Low]
Is the definition for vm_alloc_page_table_pages() missing from the tree?
Calling this in __vm_alloc_pt() leads to an undefined function build failure.

> +	} else {
> +		struct vm_pt_frag_cache *pt_frag_cache;

[ ... ]

> +void kvm_arch_vm_post_create(struct kvm_vm *vm, unsigned int nr_vcpus)
> +{
> +	size_t excp_size = __interrupts_end - __interrupts_start;
> +	size_t excp_pages = (excp_size + vm->page_size - 1) / vm->page_size;
> +	gpa_t excp_paddr;
> +	void *mem;
> +
> +	/*
> +	 * Guest interrupts are taken in real-mode, i.e. the vectors must live
> +	 * at gPA 0.  Use the innermost allocator to override the minimum GPA
> +	 * that MEM_REGION_DATA would otherwise impose.
> +	 */
> +	excp_paddr = ____vm_phy_pages_alloc(vm, excp_pages, 0,
> +					    vm->memslots[MEM_REGION_DATA],
> +					    vm_arch_has_protected_memory(vm),
> +					    false);

[Severity: Low]
Is the definition for ____vm_phy_pages_alloc() with 6 arguments missing?
Calling this undefined function in kvm_arch_vm_post_create() causes a build
failure when compiling the powerpc KVM selftests.

> +
> +	TEST_ASSERT(excp_paddr == 0,
> +		    "Interrupt vectors not allocated at gPA address 0: (0x%lx)",
> +		    excp_paddr);
> +
> +	mem = addr_gpa2hva(vm, excp_paddr);
> +	memcpy(mem, __interrupts_start, excp_size);
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/cover.1789097569.git.ritesh.list@gmail.com?part=1