Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   David Stevens <stevensd@chromium.org>
To:     Marc Zyngier <maz@kernel.org>,
        Sean Christopherson <seanjc@google.com>
Cc:     Oliver Upton <oliver.upton@linux.dev>,
        Paolo Bonzini <pbonzini@redhat.com>,
        linux-arm-kernel@lists.infradead.org, kvmarm@lists.cs.columbia.edu,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        David Stevens <stevensd@chromium.org>
Subject: [PATCH v6 0/4] KVM: allow mapping non-refcounted pages
Date:   Thu, 30 Mar 2023 17:57:58 +0900
Message-Id: <20230330085802.2414466-1-stevensd@google.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: David Stevens <stevensd@chromium.org>

This patch series adds support for mapping VM_IO and VM_PFNMAP memory
that is backed by struct pages that aren't currently being refcounted
(e.g. tail pages of non-compound higher order allocations) into the
guest.

Our use case is virtio-gpu blob resources [1], which directly map host
graphics buffers into the guest as "vram" for the virtio-gpu device.
This feature currently does not work on systems using the amdgpu driver,
as that driver allocates non-compound higher order pages via
ttm_pool_alloc_page.

Currently, the gfn_to_pfn functions require being able to pin the target
pfn, so they fail when the pfn returned by follow_pte isn't a
ref-counted page. However, the KVM secondary MMUs do not require that
the pfn be pinned, since they are integrated with the mmu notifier API.
This series adds a new set of gfn_to_pfn_noref functions which parallel
the gfn_to_pfn functions but do not pin the pfn. The new functions
return the page from gup if it was present, so callers can use it and
call put_page when done.

This series updates x86 and arm64 secondary MMUs to the new API. Other
MMUs can likely be updated without too much difficulty, but I am not
familiar with them and have no way to test them. On the other hand,
updating the rest of KVM would require replacing all usages of
kvm_vcpu_map with the gfn_to_pfn_cache, which is not at all easy [2].

[1]
https://patchwork.kernel.org/project/dri-devel/cover/20200814024000.2485-1-gurchetansingh@chromium.org/
[2] https://lore.kernel.org/all/ZBEEQtmtNPaEqU1i@google.com/

v5 -> v6:
 - rebase on kvm next branch
 - rename gfn_to_pfn_page to gfn_to_pfn_noref
 - fix uninitialized outparam in error case of __kvm_faultin_pfn
 - add kvm_release_pfn_noref_clean for releasing pfn/page pair
v4 -> v5:
 - rebase on kvm next branch again
v3 -> v4:
 - rebase on kvm next branch again
 - Add some more context to a comment in ensure_pfn_ref
v2 -> v3:
 - rebase on kvm next branch
v1 -> v2:
 - Introduce new gfn_to_pfn_page functions instead of modifying the
   behavior of existing gfn_to_pfn functions, to make the change less
   invasive.
 - Drop changes to mmu_audit.c
 - Include Nicholas Piggin's patch to avoid corrupting refcount in the
   follow_pte case, and use it in depreciated gfn_to_pfn functions.
 - Rebase on kvm/next

David Stevens (4):
  KVM: mmu: introduce new gfn_to_pfn_noref functions
  KVM: x86/mmu: use gfn_to_pfn_noref
  KVM: arm64/mmu: use gfn_to_pfn_noref
  KVM: mmu: remove over-aggressive warnings

 arch/arm64/kvm/mmu.c            |  21 ++--
 arch/x86/kvm/mmu/mmu.c          |  29 ++---
 arch/x86/kvm/mmu/mmu_internal.h |   1 +
 arch/x86/kvm/mmu/paging_tmpl.h  |   7 +-
 arch/x86/kvm/x86.c              |   5 +-
 include/linux/kvm_host.h        |  18 +++
 virt/kvm/kvm_main.c             | 214 +++++++++++++++++++++++---------
 virt/kvm/kvm_mm.h               |   6 +-
 virt/kvm/pfncache.c             |  12 +-
 9 files changed, 220 insertions(+), 93 deletions(-)

-- 
2.40.0.348.gf938b09366-goog