Commit b439eb8a authored by Sean Christopherson's avatar Sean Christopherson Committed by Paolo Bonzini
Browse files

Revert "KVM: SVM: Skip WRMSR fastpath on VM-Exit if next RIP isn't valid" 



Now that handle_fastpath_set_msr_irqoff() acquires kvm->srcu, i.e. allows
dereferencing memslots during WRMSR emulation, drop the requirement that
"next RIP" is valid.  In hindsight, acquiring kvm->srcu would have been a
better fix than avoiding the pastpath, but at the time it was thought that
accessing SRCU-protected data in the fastpath was a one-off edge case.

This reverts commit 5c30e810.

Signed-off-by: default avatarSean Christopherson <seanjc@google.com>
Message-Id: <20230721224337.2335137-3-seanjc@google.com>
Signed-off-by: default avatarPaolo Bonzini <pbonzini@redhat.com>
parent 3f2739bd

arch/x86/kvm/svm/svm.c

+2 −8
Original line number Diff line number Diff line
@@ -3986,14 +3986,8 @@ static int svm_vcpu_pre_run(struct kvm_vcpu *vcpu) 


static fastpath_t svm_exit_handlers_fastpath(struct kvm_vcpu *vcpu)

{

	struct vmcb_control_area *control = &to_svm(vcpu)->vmcb->control;



	/*

	 * Note, the next RIP must be provided as SRCU isn't held, i.e. KVM

	 * can't read guest memory (dereference memslots) to decode the WRMSR.

	 */

	if (control->exit_code == SVM_EXIT_MSR && control->exit_info_1 &&

	    nrips && control->next_rip)

	if (to_svm(vcpu)->vmcb->control.exit_code == SVM_EXIT_MSR &&

	    to_svm(vcpu)->vmcb->control.exit_info_1)

		return handle_fastpath_set_msr_irqoff(vcpu);



	return EXIT_FASTPATH_NONE;