diff mbox series

[v2,04/12] KVM: selftests: Require GPA to be aligned when backed by hugepages

Message ID 20211111000310.1435032-5-dmatlack@google.com (mailing list archive)
State New, archived
Headers show
Series KVM: selftests: Hugepage fixes and cleanups | expand

Commit Message

David Matlack Nov. 11, 2021, 12:03 a.m. UTC 
From: Sean Christopherson <seanjc@google.com>

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 <bgardon@google.com>
Cc: Yanan Wang <wangyanan55@huawei.com>
Cc: Andrew Jones <drjones@redhat.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Aaron Lewis <aaronlewis@google.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
[Used get_backing_src_pagesz() to determine alignment dynamically.]
Signed-off-by: David Matlack <dmatlack@google.com>
---
 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(-)

Comments

Ben Gardon Nov. 11, 2021, 5:49 p.m. UTC | #1 
On Wed, Nov 10, 2021 at 4:03 PM David Matlack <dmatlack@google.com> wrote:
>
> From: Sean Christopherson <seanjc@google.com>
>
> 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 <bgardon@google.com>
> Cc: Yanan Wang <wangyanan55@huawei.com>
> Cc: Andrew Jones <drjones@redhat.com>
> Cc: Peter Xu <peterx@redhat.com>
> Cc: Aaron Lewis <aaronlewis@google.com>
> Signed-off-by: Sean Christopherson <seanjc@google.com>
> [Used get_backing_src_pagesz() to determine alignment dynamically.]
> Signed-off-by: David Matlack <dmatlack@google.com>
> ---
>  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(-)
>
> 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;

Is this comment correct? I wouldn't expect the guest page size to
determine the host dirty logging granularity.

>
>         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);
> --
> 2.34.0.rc1.387.gb447b232ab-goog
>
Sean Christopherson Nov. 11, 2021, 7:21 p.m. UTC | #2 
On Thu, Nov 11, 2021, Ben Gardon wrote:
> On Wed, Nov 10, 2021 at 4:03 PM David Matlack <dmatlack@google.com> wrote:
> >
> > From: Sean Christopherson <seanjc@google.com>
> >
> > 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 <bgardon@google.com>
> > Cc: Yanan Wang <wangyanan55@huawei.com>
> > Cc: Andrew Jones <drjones@redhat.com>
> > Cc: Peter Xu <peterx@redhat.com>
> > Cc: Aaron Lewis <aaronlewis@google.com>
> > Signed-off-by: Sean Christopherson <seanjc@google.com>
> > [Used get_backing_src_pagesz() to determine alignment dynamically.]
> > Signed-off-by: David Matlack <dmatlack@google.com>
> > ---
> >  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(-)
> >
> > 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;
> 
> Is this comment correct? I wouldn't expect the guest page size to
> determine the host dirty logging granularity.

"guest page size" is a bit of a misnomer.  It's not the page size of the guest's
page tables, rather it's the non-huge page size of the PTEs that KVM uses to map
guest memory.  That info is exposed to the guest so that the guest and host agree
on the stride.

> >         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);
> > --
> > 2.34.0.rc1.387.gb447b232ab-goog
> >
diff mbox series

Patch

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);