[PATCH] KVM: selftests: Drop stale exact flags check in private_mem_kvm_exits_test

[PATCH] KVM: selftests: Drop stale exact flags check in private_mem_kvm_exits_test

[David Woodhouse]

From: David Woodhouse <dwmw@amazon.co.uk>

Commit 7d0afe669374 ("KVM: selftests: Take into account mixed memory
fault flags") relaxed the memory_fault.flags checks so that they only
test for KVM_MEMORY_EXIT_FLAG_PRIVATE, since KVM may report the access
type alongside it. In test_private_access_memslot_deleted() it added
the relaxed check but left the exact match above it, and with the
READ flag now reported the test fails for me there:

  x86/private_mem_kvm_exits_test.c:93: vcpu->run->memory_fault.flags == KVM_MEMORY_EXIT_FLAG_PRIVATE
  0x9 != 0x8 (vcpu->run->memory_fault.flags != KVM_MEMORY_EXIT_FLAG_PRIVATE)

Drop the exact match, leaving the check that the fault was private.

Fixes: 7d0afe669374 ("KVM: selftests: Take into account mixed memory fault flags")
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
---
 tools/testing/selftests/kvm/x86/private_mem_kvm_exits_test.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tools/testing/selftests/kvm/x86/private_mem_kvm_exits_test.c b/tools/testing/selftests/kvm/x86/private_mem_kvm_exits_test.c
index ed8edde9b718..e07f4972ab7e 100644
--- a/tools/testing/selftests/kvm/x86/private_mem_kvm_exits_test.c
+++ b/tools/testing/selftests/kvm/x86/private_mem_kvm_exits_test.c
@@ -90,7 +90,6 @@ static void test_private_access_memslot_deleted(void)
 		TEST_ASSERT_EQ(vcpu->run->mmio.is_write, false);
 	} else {
 		TEST_ASSERT_EQ(exit_reason, KVM_EXIT_MEMORY_FAULT);
-		TEST_ASSERT_EQ(vcpu->run->memory_fault.flags, KVM_MEMORY_EXIT_FLAG_PRIVATE);
 		TEST_ASSERT(vcpu->run->memory_fault.flags & KVM_MEMORY_EXIT_FLAG_PRIVATE,
 			    "Memory fault didn't occur on a private memory access");
 		TEST_ASSERT_EQ(vcpu->run->memory_fault.gpa, EXITS_TEST_GPA);

base-commit: e71c4e83c03926c300f6c43e79b783d47ba8b588
-- 
2.50.1

Re: [PATCH] KVM: selftests: Drop stale exact flags check in private_mem_kvm_exits_test

[Sean Christopherson]

On Thu, Nov 12, 2026, David Woodhouse wrote:
> From: David Woodhouse <dwmw@amazon.co.uk>
> 
> Commit 7d0afe669374 ("KVM: selftests: Take into account mixed memory
> fault flags") relaxed the memory_fault.flags checks so that they only
> test for KVM_MEMORY_EXIT_FLAG_PRIVATE, since KVM may report the access
> type alongside it. In test_private_access_memslot_deleted() it added
> the relaxed check but left the exact match above it, and with the
> READ flag now reported the test fails for me there:

Ah, this is against kvm/kvm-memprot-v4.

Paolo, what are your plans with that branch?  Sashiko had valid complaints, e.g.
against patch 1[*], so I assume the branch won't be merged as-is.

[*] https://lore.kernel.org/all/sashiko-outbox-193903@kernel.org

Re: [PATCH] KVM: selftests: Drop stale exact flags check in private_mem_kvm_exits_test

[David Woodhouse]

On 12 November 2026 10:22:38 GMT-07:00, Sean Christopherson <seanjc@google.com> wrote:
>On Thu, Nov 12, 2026, David Woodhouse wrote:
>> From: David Woodhouse <dwmw@amazon.co.uk>
>> 
>> Commit 7d0afe669374 ("KVM: selftests: Take into account mixed memory
>> fault flags") relaxed the memory_fault.flags checks so that they only
>> test for KVM_MEMORY_EXIT_FLAG_PRIVATE, since KVM may report the access
>> type alongside it. In test_private_access_memslot_deleted() it added
>> the relaxed check but left the exact match above it, and with the
>> READ flag now reported the test fails for me there:
>
>Ah, this is against kvm/kvm-memprot-v4.
>
>Paolo, what are your plans with that branch?  Sashiko had valid complaints, e.g.
>against patch 1[*], so I assume the branch won't be merged as-is.
>
>[*] https://lore.kernel.org/all/sashiko-outbox-193903@kernel.org

Yeah, I have a whole ASCII art diagram for VSM of the various branches it depends on, all of which need to come together.

Not sure how it'll survive being pasted on my phone...

```
  *   e7bf7087c71e (vsm-v2, vsm2/rollup) Merge branch 'vsm2/translate-probe-fix' into vsm2/rollup
  |\
  | * 2041e48469c9 (vsm2/translate-probe-fix) KVM: selftests: Add a test for KVM_TRANSLATE write permission
  | * 11e4bc669f5a KVM: x86: Report the guest's write permission from KVM_TRANSLATE
  | * (base: v7.4-rc2)                                    ← postable №2, [RFC]
  |
  * 711d67d9ee59 (vsm2/vsm-extras) ⋯ 10 EXTRA commits ⋯
  |      TLB-flush-inhibit group [Nikolas×5 + 1 ours], suspended-state,
  |      hugepage selftest [Mushahid], stats, KVM_TRANSLATE allowlist
  |
  * 45e6e98c313e (vsm2/vsm-min, tag: vsm2-min1) ⋯ 32 MIN commits ⋯
  |      per-plane attrs → TSC page → TLB flush → VTL call/return →
  |      forced entry → SynIC/overlays → intercepts → SVM state → selftests
  |      [29 dwmw, 2 Mushahid incl. folds, 1 Sriram]
  |
  *   2dba8e6aa231 Merge branch 'vsm2/hcall-exit' into vsm2/vsm-min
  |\
  | * ab78e5f86ef8 (vsm2/hcall-exit) KVM: selftests: Add Hyper-V hypercall forwarding test
  | * b5db9e2e8a73 KVM: x86: hyper-v: Advertise XMM fast hypercall output
  | * 13c74a93fa3a KVM: x86: hyper-v: Let userspace fault a forwarded hypercall
  | * 365758d0df84 KVM: x86: hyper-v: Add KVM_CAP_HYPERV_FORWARD_HCALL
  | * (base: v7.4-rc2)                                    ← postable №1
  |
  *   d834e6e32d45 Merge branch 'vsm2/memattrs' into vsm2/vsm-min
  |\
  | *   2fd5c2a710a2 (vsm2/memattrs) Merge 'kvm-memprot-v4' of kvm.git
  | |\
  | | * ⋯ 45 commits, Paolo, verbatim ⋯
  | |/
  | * (base: v7.4-rc2)
  |
  * 08140f5dfd94 (vsm2/planes-uapi) ⋯ 9 commits ⋯
  |      KVM_CAP_PLANES opt-in, run-state ioctls, request fan-out,
  |      APICv inhibit, XCR0 sharing, selftests [8 dwmw, 1 Paolo-reworked]
  |
  * 52a6b958f643 (vsm2/planes-fixes) ⋯ 16 commits ⋯
  |      s390 lkp fix first; Hyper-V init, allowlist, current_vcpu/IPIv,
  |      TSC matching, CPUID [13 dwmw, 3 Mushahid]
  |
  * ea638a7ede9e (vsm2/planes) ⋯ 60 commits, Jörg, verbatim ⋯
  |
  * (v7.4-rc2)

  parked, not merged anywhere:
  * de563700e7ed (vsm2/translate2) ours ×1 (squashed probe) on ⋯ Nikolas ×16 ⋯ on vsm2/memattrs


```