linux-kernel - Re: [RFC PATCH 13/13] KVM: SVM: Use fastpath x2apic IPI emulation when #vmexit with x2AVIC

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <34f52fb38a54e22ede0f2e28c6a0ecb49bf01a68.camel@redhat.com>
Date:   Thu, 24 Feb 2022 22:12:18 +0200
From:   Maxim Levitsky <mlevitsk@...hat.com>
To:     Suravee Suthikulpanit <suravee.suthikulpanit@....com>,
        linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Cc:     pbonzini@...hat.com, seanjc@...gle.com, joro@...tes.org,
        jon.grimm@....com, wei.huang2@....com, terry.bowman@....com
Subject: Re: [RFC PATCH 13/13] KVM: SVM: Use fastpath x2apic IPI emulation
 when #vmexit with x2AVIC

On Sun, 2022-02-20 at 20:19 -0600, Suravee Suthikulpanit wrote:
> When sends IPI to a halting vCPU, the hardware generates
> avic_incomplete_ipi #vmexit with the
> AVIC_IPI_FAILURE_TARGET_NOT_RUNNING reason.
> 
> For x2AVIC, enable fastpath emulation.
> 
> Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@....com>
> ---
>  arch/x86/kvm/svm/avic.c | 2 ++
>  arch/x86/kvm/x86.c      | 3 ++-
>  arch/x86/kvm/x86.h      | 1 +
>  3 files changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kvm/svm/avic.c b/arch/x86/kvm/svm/avic.c
> index 874c89f8fd47..758a79ee7f99 100644
> --- a/arch/x86/kvm/svm/avic.c
> +++ b/arch/x86/kvm/svm/avic.c
> @@ -428,6 +428,8 @@ int avic_incomplete_ipi_interception(struct kvm_vcpu *vcpu)
>  		kvm_lapic_reg_write(apic, APIC_ICR, icrl);
>  		break;
>  	case AVIC_IPI_FAILURE_TARGET_NOT_RUNNING:
> +		handle_fastpath_set_x2apic_icr_irqoff(vcpu, svm->vmcb->control.exit_info_1);

This just doesn't seem right - it sends IPI to the target, while we just need to wake it up.
avic_kick_target_vcpus already does all of this, and it really should be optimized to avoid
going over all vcpus as it does currently.

Best regards,
	Maxim Levitsky



> +
>  		/*
>  		 * At this point, we expect that the AVIC HW has already
>  		 * set the appropriate IRR bits on the valid target
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 641044db415d..c293027c7c10 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -2008,7 +2008,7 @@ static inline bool kvm_vcpu_exit_request(struct kvm_vcpu *vcpu)
>   * from guest to host, e.g. reacquiring KVM's SRCU lock. In contrast to the
>   * other cases which must be called after interrupts are enabled on the host.
>   */
> -static int handle_fastpath_set_x2apic_icr_irqoff(struct kvm_vcpu *vcpu, u64 data)
> +int handle_fastpath_set_x2apic_icr_irqoff(struct kvm_vcpu *vcpu, u64 data)
>  {
>  	if (!lapic_in_kernel(vcpu) || !apic_x2apic_mode(vcpu->arch.apic))
>  		return 1;
> @@ -2028,6 +2028,7 @@ static int handle_fastpath_set_x2apic_icr_irqoff(struct kvm_vcpu *vcpu, u64 data
>  
>  	return 1;
>  }
> +EXPORT_SYMBOL_GPL(handle_fastpath_set_x2apic_icr_irqoff);
>  
>  static int handle_fastpath_set_tscdeadline(struct kvm_vcpu *vcpu, u64 data)
>  {
> diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h
> index 767ec7f99516..035d20f83ca6 100644
> --- a/arch/x86/kvm/x86.h
> +++ b/arch/x86/kvm/x86.h
> @@ -286,6 +286,7 @@ int x86_decode_emulated_instruction(struct kvm_vcpu *vcpu, int emulation_type,
>  int x86_emulate_instruction(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa,
>  			    int emulation_type, void *insn, int insn_len);
>  fastpath_t handle_fastpath_set_msr_irqoff(struct kvm_vcpu *vcpu);
> +int handle_fastpath_set_x2apic_icr_irqoff(struct kvm_vcpu *vcpu, u64 data);
>  
>  extern u64 host_xcr0;
>  extern u64 supported_xcr0;