Skip to content

LAM and LASS KVM Enabling

Binbin Wu kvm lkml 6 versions – Applied

Status from this message: Applied to kvm-x86 lam (for 6.8)! I skipped the LASS patches, including patch 2

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v11 Latest16 patches 10 repliesApplied
  2. v109 patches 32 repliesSuperseded
  3. v96 patches 19 repliesSuperseded
Earlier versions · 3
VersionPostedPatchesRepliesStatus
v86 patches22 repliesSuperseded
v75 patches42 repliesSuperseded
v67 patches35 repliesSuperseded

Changes in v11 from the cover letter

- A separate patch to drop non-PA bits when getting GFN for guest's PGD [Sean]
- Add a patch to remove kvm_vcpu_is_illegal_gpa() [Isaku]
- Squash CR4 LAM bit handling with the address untag for supervisor pointers. [Sean]
- Squash CR3 LAM bits handling with the address untag for user pointers. [Sean]
- Adopt KVM-governed feature framework to track "LAM enabled" as a separate optimization patch, and add the reason in patch change log. [Sean, Kai]
- Some comment modifications/additions according to reviews [Sean]

Patches in v11 Applied

#Subject and filesSizeRepliesApplied
1/16KVM: x86: Consolidate flags for __linearize()arch/x86/kvm/emulate.c arch/x86/kvm/kvm_emulate.h+15 -100 replies
2/16KVM: x86: Use a new flag for branch targetsarch/x86/kvm/emulate.c arch/x86/kvm/kvm_emulate.h+4 -21 reply
3/16KVM: x86: Add an emulation flag for implicit system accessarch/x86/kvm/kvm_emulate.h+1 -00 replies
4/16KVM: x86: Add X86EMUL_F_INVLPG and pass it in em_invlpg()arch/x86/kvm/emulate.c arch/x86/kvm/kvm_emulate.h+4 -10 replies
5/16KVM: x86/mmu: Drop non-PA bits when getting GFN for guest's PGDarch/x86/kvm/mmu/mmu.c arch/x86/kvm/mmu/mmu_internal.h arch/x86/kvm/mmu/paging_tmpl.h+3 -20 replies
6/16KVM: x86: Add & use kvm_vcpu_is_legal_cr3() to check CR3's legalityarch/x86/kvm/cpuid.h arch/x86/kvm/svm/nested.c arch/x86/kvm/vmx/nested.c +1 more+11 -60 replies
7/16KVM: x86: Remove kvm_vcpu_is_illegal_gpa()arch/x86/kvm/cpuid.h arch/x86/kvm/vmx/nested.c arch/x86/kvm/vmx/vmx.c+2 -70 replies
8/16KVM: x86: Introduce get_untagged_addr() in kvm_x86_ops and call it in emulatorarch/x86/include/asm/kvm-x86-ops.h arch/x86/include/asm/kvm_host.h arch/x86/kvm/emulate.c +2 more+17 -12 replies
9/16KVM: x86: Untag address for vmexit handlers when LAM applicablearch/x86/kvm/vmx/nested.c arch/x86/kvm/vmx/sgx.c arch/x86/kvm/vmx/vmx.c +2 more+19 -00 replies
10/16KVM: x86: Virtualize LAM for supervisor pointerarch/x86/include/asm/kvm_host.h arch/x86/kvm/vmx/vmx.c arch/x86/kvm/x86.h+43 -20 replies
11/16KVM: x86: Virtualize LAM for user pointerarch/x86/kvm/cpuid.h arch/x86/kvm/mmu.h arch/x86/kvm/vmx/vmx.c+22 -30 replies
12/16KVM: x86: Advertise and enable LAM (user and supervisor)arch/x86/kvm/cpuid.c+1 -10 replies
13/16KVM: x86: Use KVM-governed feature framework to track "LAM enabled"arch/x86/kvm/cpuid.h arch/x86/kvm/governed_features.h arch/x86/kvm/mmu.h +1 more+4 -40 replies
14/16KVM: emulator: Add emulation of LASS violation checks on linear addressarch/x86/include/asm/kvm-x86-ops.h arch/x86/include/asm/kvm_host.h arch/x86/kvm/emulate.c +2 more+29 -10 replies
15/16KVM: VMX: Virtualize LASSarch/x86/include/asm/kvm_host.h arch/x86/kvm/vmx/nested.c arch/x86/kvm/vmx/sgx.c +4 more+62 -50 replies
16/16KVM: x86: Advertise LASS CPUID to user spacearch/x86/kvm/cpuid.c+1 -10 replies

Reviews trailers per patch

6 of 16 patches have a Reviewed-by or Acked-by.

PatchChao GaoKai HuangXuelian GuoJingqi Liu
1/16R-b A-b T-b
2/16T-b
3/16T-b
4/16T-b
5/16T-b
6/16T-b
7/16T-b
8/16R-b T-b
9/16R-b T-b
10/16R-b R-b T-b
11/16R-b R-b T-b
12/16R-b R-b T-b R-b
13/16T-b
14/16T-b
15/16T-b
16/16T-b

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 v11 thread

Cover letter [PATCH v11 00/16] LAM and LASS KVM Enabling

This patch series includes KVM enabling patches for Linear-address masking (LAM) v11 and Linear Address Space Separation (LASS) v3 since the two features have overlapping prep work and concepts. Sent as a single series to reduce the probability of conflicts.

The patch series is organized as follows:
- Patch 1-4: Common prep work for both LAM and LASS.
- Patch 5-13: LAM part.
- Patch 14-16: LASS part.

Dependency:
- LAM has no other dependency.
- LASS patches depends on LASS kernel enabling patches, which are not merged yet.
https://lore.kernel.org/all/[email protected]/

==== LAM v11 ====

Linear-address masking (LAM) [1], modifies the checking that is applied to *64-bit* linear addresses, allowing software to use of the untranslated address bits for metadata and masks the metadata bits before using them as linear addresses to access memory.

When the feature is virtualized and exposed to guest, it can be used for efficient address sanitizers (ASAN) implementation and for optimizations in JITs and virtual machines.

The patch series brings LAM virtualization support in KVM.

Please review and consider applying.

LAM QEMU patch:
https://lists.gnu.org/archive/html/qemu-devel/2023-07/msg04160.html

LAM kvm-unit-tests patch:
https://lore.kernel.org/kvm/[email protected]/

--- Test ---
1. Add test cases in kvm-unit-test for LAM [2], including LAM_SUP and LAM_{U57,U48}. For supervisor pointers, the test covers CR4 LAM_SUP bits toggle, Memory/MMIO access with tagged pointer, and some special instructions (INVLPG, INVPCID, INVVPID), INVVPID cases also used to cover VMX instruction VMExit path. For user pointers, the test covers CR3 LAM bits toggle, Memory/MMIO access with tagged pointer.
MMIO cases are used to trigger instruction emulation path.
Run the unit test with both LAM feature on/off (i.e. including negative cases). Run the unit test in L1 guest with both LAM feature on/off.
2. Run Kernel LAM kselftests in guest, with both EPT=Y/N.
3. Launch a nested guest and run tests listed in 1 & 2.

All tests have passed on real machine supporting LAM.

[1] Intel ISE https://cdrdv2.intel.com/v1/dl/getContent/671368
Chapter Linear Address Masking (LAM)
[2] https://lore.kernel.org/kvm/[email protected]/

----------
Changelog

v11:
- A separate patch to drop non-PA bits when getting GFN for guest's PGD [Sean]
- Add a patch to remove kvm_vcpu_is_illegal_gpa() [Isaku]
- Squash CR4 LAM bit handling with the address untag for supervisor pointers. [Sean]
- Squash CR3 LAM bits handling with the address untag for user pointers. [Sean]
- Adopt KVM-governed feature framework to track "LAM enabled" as a separate optimization patch, and add the reason in patch change log. [Sean, Kai]
- Some comment modifications/additions according to reviews [Sean]

v10:
https://lore.kernel.org/kvm/[email protected]/

==== LASS v3 ====

Linear Address Space Separation (LASS)[1] is a new mechanism that enforces the same mode-based protections as paging, i.e. SMAP/SMEP but without traversing the paging structures. Because the protections enforced by LASS are applied before paging, "probes" by malicious software will provide no paging-based timing information.

This patch series provide a LASS KVM solution and depends on kernel enabling that can be found at [2].

--- Test ---
1. Test the basic function of LASS virtualization including LASS enumeration and enabling in guest and nested environment.
2. Run selftest with following cases:
- data access to user address space in supervisor mode
- data access to supervisor address space in user mode
- data access to linear address across space boundary
- Using KVM FEP mechanism to run test cases above
- VMX instruction execution with VMCS structure in user address space
- instruction fetch from user address space in supervisor mode
- instruction fetch from supervisor address space in user mode

All tests have passed on real machine supporting LASS.

[1] Intel ISE spec https://cdrdv2.intel.com/v1/dl/getContent/671368
Chapter Linear Address Space Separation (LASS)

[2] LASS kernel patch series
https://lore.kernel.org/all/[email protected]/

----------
Change log

v3:
1. Refine commit message [Sean/Chao Gao]
2. Enhance the implementation of LASS violation check [Sean]
3. Re-organize patch as Sean's suggestion [Sean]

v2:
https://lore.kernel.org/all/[email protected]/

Binbin Wu (10):
KVM: x86: Consolidate flags for __linearize()
KVM: x86: Use a new flag for branch targets
KVM: x86: Add an emulation flag for implicit system access
KVM: x86: Add X86EMUL_F_INVLPG and pass it in em_invlpg()
KVM: x86/mmu: Drop non-PA bits when getting GFN for guest's PGD
KVM: x86: Add & use kvm_vcpu_is_legal_cr3() to check CR3's legality
KVM: x86: Remove kvm_vcpu_is_illegal_gpa()
KVM: x86: Introduce get_untagged_addr() in kvm_x86_ops and call it in
emulator
KVM: x86: Untag address for vmexit handlers when LAM applicable
KVM: x86: Use KVM-governed feature framework to track "LAM enabled"

Robert Hoo (3):
KVM: x86: Virtualize LAM for supervisor pointer
KVM: x86: Virtualize LAM for user pointer
KVM: x86: Advertise and enable LAM (user and supervisor)

Zeng Guang (3):
KVM: emulator: Add emulation of LASS violation checks on linear
address
KVM: VMX: Virtualize LASS
KVM: x86: Advertise LASS CPUID to user space

 arch/x86/include/asm/kvm-x86-ops.h |   4 +-
 arch/x86/include/asm/kvm_host.h    |   8 ++-
 arch/x86/kvm/cpuid.c               |   4 +-
 arch/x86/kvm/cpuid.h               |  13 ++--
 arch/x86/kvm/emulate.c             |  39 +++++++----
 arch/x86/kvm/governed_features.h   |   1 +
 arch/x86/kvm/kvm_emulate.h         |  13 ++++
 arch/x86/kvm/mmu.h                 |   8 +++
 arch/x86/kvm/mmu/mmu.c             |   2 +-
 arch/x86/kvm/mmu/mmu_internal.h    |   1 +
 arch/x86/kvm/mmu/paging_tmpl.h     |   2 +-
 arch/x86/kvm/svm/nested.c          |   4 +-
 arch/x86/kvm/vmx/nested.c          |  14 ++--
 arch/x86/kvm/vmx/sgx.c             |   4 +-
 arch/x86/kvm/vmx/vmx.c             | 106 ++++++++++++++++++++++++++++-
 arch/x86/kvm/vmx/vmx.h             |   5 ++
 arch/x86/kvm/x86.c                 |  28 +++++++-
 arch/x86/kvm/x86.h                 |   4 ++
 18 files changed, 226 insertions(+), 34 deletions(-)

base-commit: 0bb80ecc33a8fb5a682236443c1e740d5c917d1d
prerequisite-patch-id: 51db36ad7156234d05f8c4004ec6a31ef609b81a

Signature · 1 line
2.25.1