Skip to contentSkip to conversation

KVM: X86: Explicitly set the 'fault.async_page_fault' value in kvm_fixup_and_inject_pf_error().

PATCH Applied Yu Zhang 3 participants 6 messages Jul 18, 2022 – Aug 5, 2022

Subsystems (3)
lore
Share & tools
lore

Yu Zhang

author
#0 lore raw

PATCH v2 1/1 KVM: X86: Explicitly set the 'fault.async_page_fault' value in kvm_fixup_and_inject_pf_error().

kvm_fixup_and_inject_pf_error() was introduced to fixup the error code( e.g., to add RSVD flag) and inject the #PF to the guest, when guest MAXPHYADDR is smaller than the host one.

When it comes to nested, L0 is expected to intercept and fix up the #PF and then inject to L2 directly if
- L2.MAXPHYADDR < L0.MAXPHYADDR and
- L1 has no intention to intercept L2's #PF (e.g., L2 and L1 have the same MAXPHYADDR value && L1 is using EPT for L2),
instead of constructing a #PF VM Exit to L1. Currently, with PFEC_MASK and PFEC_MATCH both set to 0 in vmcs02, the interception and injection may happen on all L2 #PFs.

However, failing to initialize 'fault' in kvm_fixup_and_inject_pf_error() may cause the fault.async_page_fault being NOT zeroed, and later the #PF being treated as a nested async page fault, and then being injected to L1. Instead of zeroing 'fault' at the beginning of this function, we mannually set the value of 'fault.async_page_fault', because false is the value we really expect.

Fixes 897861479c064 ("KVM: x86: Add helper functions for illegal GPA checking and page fault injection")

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=216178

---
v2:
- Set 'fault.async_page_fault' mannually to false, instead of initializing the whole structure based on Sean's suggestion.
- Commit message changes based on the code change.

---
 arch/x86/kvm/x86.c | 1 +
 1 file changed, 1 insertion(+)
arch/x86/kvm/x86.c +1 -0
index 031678eff28e..ffd8c96cfaa5 100644
@@ -12999,6 +12999,7 @@ void kvm_fixup_and_inject_pf_error(struct kvm_vcpu *vcpu, gva_t gva, u16 error_c
 		fault.error_code = error_code;
 		fault.nested_page_fault = false;
 		fault.address = gva;
+		fault.async_page_fault = false;
 	}
 	vcpu->arch.walk_mmu->inject_page_fault(vcpu, &fault);
 }
Signature · 1 line
2.25.1
maintainer Reviewed-by on v2 1/1
#1 lore raw
Quoted 20 lines · Yu Zhang

kvm_fixup_and_inject_pf_error() was introduced to fixup the error code( e.g., to add RSVD flag) and inject the #PF to the guest, when guest MAXPHYADDR is smaller than the host one.

When it comes to nested, L0 is expected to intercept and fix up the #PF and then inject to L2 directly if
- L2.MAXPHYADDR < L0.MAXPHYADDR and
- L1 has no intention to intercept L2's #PF (e.g., L2 and L1 have the same MAXPHYADDR value && L1 is using EPT for L2),
instead of constructing a #PF VM Exit to L1. Currently, with PFEC_MASK and PFEC_MATCH both set to 0 in vmcs02, the interception and injection may happen on all L2 #PFs.

However, failing to initialize 'fault' in kvm_fixup_and_inject_pf_error() may cause the fault.async_page_fault being NOT zeroed, and later the #PF being treated as a nested async page fault, and then being injected to L1. Instead of zeroing 'fault' at the beginning of this function, we mannually set the value of 'fault.async_page_fault', because false is the value we really expect.

Quoted 5 lines
Fixes 897861479c064 ("KVM: x86: Add helper functions for illegal GPA checking and page fault injection")

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=216178

No need for my SoB, I was just providing feedback. Other than that,

author
#2 lore raw
Quoted 21 lines · Sean Christopherson
Quoted 20 lines · Yu Zhang

kvm_fixup_and_inject_pf_error() was introduced to fixup the error code( e.g., to add RSVD flag) and inject the #PF to the guest, when guest MAXPHYADDR is smaller than the host one.

When it comes to nested, L0 is expected to intercept and fix up the #PF and then inject to L2 directly if
- L2.MAXPHYADDR < L0.MAXPHYADDR and
- L1 has no intention to intercept L2's #PF (e.g., L2 and L1 have the same MAXPHYADDR value && L1 is using EPT for L2),
instead of constructing a #PF VM Exit to L1. Currently, with PFEC_MASK and PFEC_MATCH both set to 0 in vmcs02, the interception and injection may happen on all L2 #PFs.

However, failing to initialize 'fault' in kvm_fixup_and_inject_pf_error() may cause the fault.async_page_fault being NOT zeroed, and later the #PF being treated as a nested async page fault, and then being injected to L1. Instead of zeroing 'fault' at the beginning of this function, we mannually set the value of 'fault.async_page_fault', because false is the value we really expect.

Quoted 8 lines
Quoted 5 lines
Fixes 897861479c064 ("KVM: x86: Add helper functions for illegal GPA checking and page fault injection")

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=216178

No need for my SoB, I was just providing feedback. Other than that,

Thanks! It's a very detailed suggestion. :)

Quoted 2 lines

@Paolo Any comment on this fix, and on the test case change(https://www.spinics.net/lists/kvm/msg283600.html)? Thanks!

B.R.
Yu

author
#3 lore raw
Quoted 32 lines · Yu Zhang
Quoted 21 lines · Sean Christopherson
Quoted 20 lines · Yu Zhang

kvm_fixup_and_inject_pf_error() was introduced to fixup the error code( e.g., to add RSVD flag) and inject the #PF to the guest, when guest MAXPHYADDR is smaller than the host one.

When it comes to nested, L0 is expected to intercept and fix up the #PF and then inject to L2 directly if
- L2.MAXPHYADDR < L0.MAXPHYADDR and
- L1 has no intention to intercept L2's #PF (e.g., L2 and L1 have the same MAXPHYADDR value && L1 is using EPT for L2),
instead of constructing a #PF VM Exit to L1. Currently, with PFEC_MASK and PFEC_MATCH both set to 0 in vmcs02, the interception and injection may happen on all L2 #PFs.

However, failing to initialize 'fault' in kvm_fixup_and_inject_pf_error() may cause the fault.async_page_fault being NOT zeroed, and later the #PF being treated as a nested async page fault, and then being injected to L1. Instead of zeroing 'fault' at the beginning of this function, we mannually set the value of 'fault.async_page_fault', because false is the value we really expect.

Quoted 8 lines
Quoted 5 lines
Fixes 897861479c064 ("KVM: x86: Add helper functions for illegal GPA checking and page fault injection")

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=216178

No need for my SoB, I was just providing feedback. Other than that,

Thanks! It's a very detailed suggestion. :)

Quoted 7 lines
Quoted 2 lines

@Paolo Any comment on this fix, and on the test case change(https://www.spinics.net/lists/kvm/msg283600.html)? Thanks!

B.R.
Yu

Ping... Or should I send another version? Thanks!

B.R.
Yu

#4 lore raw
Quoted 36 lines · Yu Zhang
Quoted 35 lines · Yu Zhang
Quoted 21 lines · Sean Christopherson
Quoted 20 lines · Yu Zhang

kvm_fixup_and_inject_pf_error() was introduced to fixup the error code( e.g., to add RSVD flag) and inject the #PF to the guest, when guest MAXPHYADDR is smaller than the host one.

When it comes to nested, L0 is expected to intercept and fix up the #PF and then inject to L2 directly if
- L2.MAXPHYADDR < L0.MAXPHYADDR and
- L1 has no intention to intercept L2's #PF (e.g., L2 and L1 have the same MAXPHYADDR value && L1 is using EPT for L2),
instead of constructing a #PF VM Exit to L1. Currently, with PFEC_MASK and PFEC_MATCH both set to 0 in vmcs02, the interception and injection may happen on all L2 #PFs.

However, failing to initialize 'fault' in kvm_fixup_and_inject_pf_error() may cause the fault.async_page_fault being NOT zeroed, and later the #PF being treated as a nested async page fault, and then being injected to L1. Instead of zeroing 'fault' at the beginning of this function, we mannually set the value of 'fault.async_page_fault', because false is the value we really expect.

Quoted 8 lines
Quoted 5 lines
Fixes 897861479c064 ("KVM: x86: Add helper functions for illegal GPA checking and page fault injection")

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=216178

No need for my SoB, I was just providing feedback. Other than that,

Thanks! It's a very detailed suggestion. :)

Quoted 2 lines
Quoted 6 lines
Quoted 4 lines

@Paolo Any comment on this fix, and on the test case change(https://www.spinics.net/lists/kvm/msg283600.html)? Thanks!

B.R.
Yu

Ping... Or should I send another version? Thanks!

Shouldn't need to send a new version.

Paolo, can you grab this for 6.0-rc1?