[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241213195711.316050-1-pbonzini@redhat.com>
Date: Fri, 13 Dec 2024 14:56:53 -0500
From: Paolo Bonzini <pbonzini@...hat.com>
To: linux-kernel@...r.kernel.org,
kvm@...r.kernel.org
Cc: yan.y.zhao@...el.com,
isaku.yamahata@...el.com,
binbin.wu@...ux.intel.com,
rick.p.edgecombe@...el.com
Subject: [PATCH v5 00/18] TDX MMU prep series part 1
Hi,
this is the essentially final version of the TDX MMU prep series, focusing
on supporting TDX's separation of EPT into a direct part (for shared pages)
and a part that is managed by the TDX module and cached (into a "mirror"
EPT) by KVM.
The changes from v4 (https://patchew.org/linux/20240718211230.1492011-1-rick.p.edgecombe@intel.com/)
are minor:
- patch 7: kvm_tdp_mmu_handle_gfn is now __kvm_tdp_mmu_age_gfn_range
- patch 7: zap_collapsible_spte_range is now split into
tdp_mmu_make_huge_spte and recover_huge_pages_range
- patch 10/12: KVM_INVALID_ROOTS used to mean "walk all invalid roots";
now it means "walk *also* invalid roots of the kind (direct/mirror)
specified by the other bits. This is closer in meaning to the
existing code, as kvm_tdp_mmu_unmap_gfn_range() will then operate
only on direct or only on mirror pages depending on the path that
caused the invalidation (guest_memfd vs. MMU notifier)
- patch 13: adjust commit message due to change from kvm_tdp_mmu_handle_gfn
to __kvm_tdp_mmu_age_gfn_range; "or" KVM_INVALID_ROOTS into the
"types" variable in kvm_tdp_mmu_unmap_gfn_range, otherwise the loop
would not affect invalid roots. This is the problematic code from v4:
- __for_each_tdp_mmu_root_yield_safe(kvm, root, range->slot->as_id, KVM_ALL_ROOTS)
+ types = kvm_gfn_range_filter_to_root_types(kvm, range->attr_filter);
+
+ __for_each_tdp_mmu_root_yield_safe(kvm, root, range->slot->as_id, types)
and here is it in v5
+ types = kvm_gfn_range_filter_to_root_types(kvm, range->attr_filter) | KVM_INVALID_ROOTS;
+
+ __for_each_tdp_mmu_root_yield_safe(kvm, root, range->slot->as_id, types)
- patch 14: tdp_mmu_zap_spte_atomic() disappeared in commit 35ef80eb29ab
("KVM: x86/mmu: Batch TLB flushes when zapping collapsible TDP MMU SPTEs", 2024-10-30)
- patch 18: context changes due to kvm_release_pfn_clean -> kvm_mmu_finish_page_fault
Thanks,
Paolo
Isaku Yamahata (12):
KVM: Add member to struct kvm_gfn_range for target alias
KVM: x86/mmu: Add an external pointer to struct kvm_mmu_page
KVM: x86/mmu: Add an is_mirror member for union kvm_mmu_page_role
KVM: x86/tdp_mmu: Take struct kvm in iter loops
KVM: x86/mmu: Support GFN direct bits
KVM: x86/tdp_mmu: Extract root invalid check from tdx_mmu_next_root()
KVM: x86/tdp_mmu: Introduce KVM MMU root types to specify page table
type
KVM: x86/tdp_mmu: Take root in tdp_mmu_for_each_pte()
KVM: x86/tdp_mmu: Support mirror root for TDP MMU
KVM: x86/tdp_mmu: Propagate building mirror page tables
KVM: x86/tdp_mmu: Propagate tearing down mirror page tables
KVM: x86/tdp_mmu: Take root types for
kvm_tdp_mmu_invalidate_all_roots()
Paolo Bonzini (1):
KVM: x86/tdp_mmu: Propagate attr_filter to MMU notifier callbacks
Rick Edgecombe (5):
KVM: x86/mmu: Zap invalid roots with mmu_lock holding for write at
uninit
KVM: x86: Add a VM type define for TDX
KVM: x86/mmu: Make kvm_tdp_mmu_alloc_root() return void
KVM: x86/tdp_mmu: Don't zap valid mirror roots in
kvm_tdp_mmu_zap_all()
KVM: x86/mmu: Prevent aliased memslot GFNs
arch/x86/include/asm/kvm-x86-ops.h | 4 +
arch/x86/include/asm/kvm_host.h | 26 ++-
arch/x86/include/uapi/asm/kvm.h | 1 +
arch/x86/kvm/mmu.h | 31 +++
arch/x86/kvm/mmu/mmu.c | 50 ++++-
arch/x86/kvm/mmu/mmu_internal.h | 64 +++++-
arch/x86/kvm/mmu/spte.h | 5 +
arch/x86/kvm/mmu/tdp_iter.c | 10 +-
arch/x86/kvm/mmu/tdp_iter.h | 21 +-
arch/x86/kvm/mmu/tdp_mmu.c | 323 ++++++++++++++++++++++-------
arch/x86/kvm/mmu/tdp_mmu.h | 51 ++++-
arch/x86/kvm/x86.c | 3 +
include/linux/kvm_host.h | 6 +
virt/kvm/guest_memfd.c | 2 +
virt/kvm/kvm_main.c | 14 ++
15 files changed, 506 insertions(+), 105 deletions(-)
--
2.43.5
Powered by blists - more mailing lists