Skip to content

TDX MMU refactors

Rick Edgecombe kvm lkml 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest17 patches 49 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/17x86/tdx: Use pg_level in TDX APIs, not the TDX-Module's 0-based levelarch/x86/include/asm/tdx.h arch/x86/kvm/vmx/tdx.c arch/x86/virt/vmx/tdx/tdx.c+26 -250 replies
2/17KVM: x86/mmu: Update iter->old_spte if cmpxchg64 on mirror SPTE "fails"arch/x86/kvm/mmu/tdp_mmu.c+5 -55 replies
3/17KVM: TDX: Account all non-transient page allocations for per-TD structuresarch/x86/kvm/vmx/tdx.c+7 -60 replies
4/17KVM: x86: Make "external SPTE" ops that can fail RET0 static callsarch/x86/include/asm/kvm-x86-ops.h+3 -30 replies
5/17KVM: x86/tdp_mmu: Drop zapping KVM_BUG_ON() set_external_spte_present()arch/x86/kvm/mmu/tdp_mmu.c+0 -100 replies
6/17KVM: x86/tdp_mmu: Morph the !is_frozen_spte() check into a KVM_MMU_WARN_ON()arch/x86/kvm/mmu/tdp_mmu.c+7 -15 replies
7/17KVM: x86/tdp_mmu: Centralize updates to present external PTEsarch/x86/kvm/mmu/tdp_mmu.c+88 -6219 replies
8/17KVM: TDX: Drop kvm_x86_ops.link_external_spt(), use .set_external_spte() for allarch/x86/include/asm/kvm-x86-ops.h arch/x86/include/asm/kvm_host.h arch/x86/kvm/mmu/tdp_mmu.c +1 more+43 -531 reply
9/17KVM: TDX: Add helper to handle mapping leaf SPTE into S-EPTarch/x86/kvm/vmx/tdx.c+13 -82 replies
10/17KVM: TDX: Move set_external_spte_present() assert into TDX codearch/x86/kvm/mmu/tdp_mmu.c arch/x86/kvm/vmx/tdx.c+2 -33 replies
11/17KVM: x86/mmu: Fold set_external_spte_present() into its sole callerarch/x86/kvm/mmu/tdp_mmu.c+1 -82 replies
12/17KVM: x86/mmu: Plumb the old_spte into kvm_x86_ops.set_external_spte()arch/x86/include/asm/kvm_host.h arch/x86/kvm/mmu/tdp_mmu.c arch/x86/kvm/vmx/tdx.c+13 -120 replies
13/17KVM: TDX: Hoist tdx_sept_remove_private_spte() above set_private_spte()arch/x86/kvm/vmx/tdx.c+14 -142 replies
14/17KVM: x86/mmu: Remove KVM_BUG_ON() that checks lock when removing PTsarch/x86/kvm/mmu/tdp_mmu.c+1 -34 replies
15/17KVM: TDX: Handle removal of leaf SPTEs in .set_private_spte()arch/x86/include/asm/kvm-x86-ops.h arch/x86/include/asm/kvm_host.h arch/x86/kvm/mmu/tdp_mmu.c +1 more+14 -450 replies
16/17KVM: x86: Move error handling inside free_external_spt()arch/x86/include/asm/kvm-x86-ops.h arch/x86/include/asm/kvm_host.h arch/x86/kvm/mmu/tdp_mmu.c +1 more+15 -281 reply
17/17KVM: TDX: Move external page table freeing to TDX codearch/x86/kvm/mmu/tdp_mmu.c arch/x86/kvm/vmx/tdx.c+21 -65 replies

Reviews trailers per patch

1 of 17 patches has a Reviewed-by or Acked-by.

PatchKiryl ShutsemauKai HuangRick EdgecombeSean Christopherson
1/17A-b R-b T-b R-b T-b
2/17
3/17
4/17
5/17
6/17
7/17Not-yet-Signed-off-by
8/17
9/17
10/17
11/17
12/17
13/17
14/17
15/17
16/17
17/17Not-yet-Signed-off-by

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes. Scroll the table sideways for all 4 reviewers.

Read the v1 thread

Cover letter [PATCH 00/17] TDX MMU refactors

Hi,

This is all the MMU refactors extracted from the discussion on Sean’s DPAMT/Huge page combined series [0]. I was thinking that since they are valid cleanups in their own way, we could maybe get them upstream ahead of time. Yan and I rebased DPAMT/huge pages on this series and made sure it all still fits together.

It’s not suggested for merging yet. We should go over it again for a more thorough bug check with everything put together. Sean/Paolo please feel free to ignore for now, except to maybe comment if you have any objections to merging it earlier as a cleanup series.

Changes from on-list discussion:
- Go back to free_external_spt() name. Since free_external_sp() was dropped from the changes there was no similarly named function to confuse.
- Suggestions around dropping or moving KVM_BUG_ON/WARNs turned into patches.

Another possible point of discussion:
- Is the last patch needed and does it fit in to a general refactor/cleanup? (Yan)

Since some of the patches were created from diffs posted by Sean without SOB, I left them not signed off for now.

Based on kvm-x86-next-2026.03.13, testing was our usual CI suite. AI review is pending. We'll work through any Sashiko feedback on the list.

Thanks,
Rick

[0] https://lore.kernel.org/kvm/[email protected]/

Rick Edgecombe (4):
KVM: x86/tdp_mmu: Drop zapping KVM_BUG_ON()
set_external_spte_present()
KVM: x86/tdp_mmu: Morph the !is_frozen_spte() check into a

    KVM_MMU_WARN_ON()

KVM: TDX: Move set_external_spte_present() assert into TDX code
KVM: x86/mmu: Remove KVM_BUG_ON() that checks lock when removing PTs

Sean Christopherson (13):
x86/tdx: Use pg_level in TDX APIs, not the TDX-Module's 0-based level
KVM: x86/mmu: Update iter->old_spte if cmpxchg64 on mirror SPTE

    "fails"

KVM: TDX: Account all non-transient page allocations for per-TD
structures
KVM: x86: Make "external SPTE" ops that can fail RET0 static calls
KVM: x86/tdp_mmu: Centralize updates to present external PTEs
KVM: TDX: Drop kvm_x86_ops.link_external_spt(), use
.set_external_spte() for all
KVM: TDX: Add helper to handle mapping leaf SPTE into S-EPT
KVM: x86/mmu: Fold set_external_spte_present() into its sole caller
KVM: x86/mmu: Plumb the old_spte into kvm_x86_ops.set_external_spte()
KVM: TDX: Hoist tdx_sept_remove_private_spte() above
set_private_spte()
KVM: TDX: Handle removal of leaf SPTEs in .set_private_spte()
KVM: x86: Move error handling inside free_external_spt()
KVM: TDX: Move external page table freeing to TDX code

 arch/x86/include/asm/kvm-x86-ops.h |   4 +-
 arch/x86/include/asm/kvm_host.h    |  13 +-
 arch/x86/include/asm/tdx.h         |  14 +-
 arch/x86/kvm/mmu/tdp_mmu.c         | 253 ++++++++++++-----------------
 arch/x86/kvm/vmx/tdx.c             | 147 ++++++++++-------
 arch/x86/virt/vmx/tdx/tdx.c        |  26 ++-
 6 files changed, 219 insertions(+), 238 deletions(-)
Signature · 1 line
2.53.0