[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZB3eU8W9D3c50Kak@google.com>
Date: Fri, 24 Mar 2023 10:30:59 -0700
From: Sean Christopherson <seanjc@...gle.com>
To: Peter Gonda <pgonda@...gle.com>
Cc: kvm@...r.kernel.org, linux-kernel@...r.kernel.org,
Michael Roth <michael.roth@....com>,
Paolo Bonzini <pbonzini@...hat.com>,
Vishal Annapurve <vannapurve@...gle.com>,
Ackerley Tng <ackerleytng@...gle.com>,
Andrew Jones <andrew.jones@...ux.dev>
Subject: Re: [PATCH V6 4/7] KVM: selftests: add support for protected
vm_vaddr_* allocations
On Tue, Jan 10, 2023, Peter Gonda wrote:
> From: Michael Roth <michael.roth@....com>
>
> Test programs may wish to allocate shared vaddrs for things like
> sharing memory with the guest. Since protected vms will have their
> memory encrypted by default an interface is needed to explicitly
> request shared pages.
>
> Implement this by splitting the common code out from vm_vaddr_alloc()
> and introducing a new vm_vaddr_alloc_shared().
>
> Cc: Paolo Bonzini <pbonzini@...hat.com>
> Cc: Sean Christopherson <seanjc@...gle.com>
> Cc: Vishal Annapurve <vannapurve@...gle.com>
> Cc: Ackerley Tng <ackerleytng@...gle.com>
> cc: Andrew Jones <andrew.jones@...ux.dev>
> Signed-off-by: Michael Roth <michael.roth@....com>
> Signed-off-by: Peter Gonda <pgonda@...gle.com>
> ---
> .../selftests/kvm/include/kvm_util_base.h | 1 +
> tools/testing/selftests/kvm/lib/kvm_util.c | 21 +++++++++++++++----
> 2 files changed, 18 insertions(+), 4 deletions(-)
>
> diff --git a/tools/testing/selftests/kvm/include/kvm_util_base.h b/tools/testing/selftests/kvm/include/kvm_util_base.h
> index f84d7777d5ca..5f3150ecfbbf 100644
> --- a/tools/testing/selftests/kvm/include/kvm_util_base.h
> +++ b/tools/testing/selftests/kvm/include/kvm_util_base.h
> @@ -435,6 +435,7 @@ vm_vaddr_t vm_vaddr_unused_gap(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_mi
> vm_vaddr_t vm_vaddr_alloc(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min);
> vm_vaddr_t __vm_vaddr_alloc(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min,
> enum kvm_mem_region_type type);
> +vm_vaddr_t vm_vaddr_alloc_shared(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min);
Wrap.
> vm_vaddr_t vm_vaddr_alloc_pages(struct kvm_vm *vm, int nr_pages);
> vm_vaddr_t __vm_vaddr_alloc_page(struct kvm_vm *vm,
> enum kvm_mem_region_type type);
> diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c
> index ba771c2d949d..0d0a7ad7632d 100644
> --- a/tools/testing/selftests/kvm/lib/kvm_util.c
> +++ b/tools/testing/selftests/kvm/lib/kvm_util.c
> @@ -1305,15 +1305,17 @@ vm_vaddr_t vm_vaddr_unused_gap(struct kvm_vm *vm, size_t sz,
> return pgidx_start * vm->page_size;
> }
>
> -vm_vaddr_t __vm_vaddr_alloc(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min,
> - enum kvm_mem_region_type type)
> +static vm_vaddr_t ____vm_vaddr_alloc(struct kvm_vm *vm, size_t sz,
> + vm_vaddr_t vaddr_min,
> + enum kvm_mem_region_type type,
> + bool encrypt)
s/encrypt/protected, or maybe mark_protected so that it's clear that it's a command.
> {
> uint64_t pages = (sz >> vm->page_shift) + ((sz % vm->page_size) != 0);
>
> virt_pgd_alloc(vm);
> - vm_paddr_t paddr = vm_phy_pages_alloc(vm, pages,
> + vm_paddr_t paddr = _vm_phy_pages_alloc(vm, pages,
> KVM_UTIL_MIN_PFN * vm->page_size,
> - vm->memslots[type]);
> + vm->memslots[type], encrypt);
>
> /*
> * Find an unused range of virtual page addresses of at least
> @@ -1333,6 +1335,17 @@ vm_vaddr_t __vm_vaddr_alloc(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min,
> return vaddr_start;
> }
>
> +vm_vaddr_t __vm_vaddr_alloc(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min,
> + enum kvm_mem_region_type type)
> +{
> + return ____vm_vaddr_alloc(vm, sz, vaddr_min, type, vm->protected);
> +}
> +
> +vm_vaddr_t vm_vaddr_alloc_shared(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min)
> +{
> + return ____vm_vaddr_alloc(vm, sz, vaddr_min, MEM_REGION_TEST_DATA, false);
This shouldn't be MEM_REGION_TEST_DATA, because then the ucall patches changes
from MEM_REGION_DATA to MEM_REGION_TEST_DATA, which I suspect will break ARM's
page_fault_test. So, we probably need to have vm_vaddr_alloc_shared() be a true
mirror of the non-shared version and take @type.
Powered by blists - more mailing lists