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] [day] [month] [year] [list]
Date:   Fri, 23 Dec 2016 13:08:55 +0100
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Cc:     rkrcmar@...hat.com, rkagan@...tuozzo.com, dplotnikov@...tuozzo.com
Subject: Re: [PATCH 2/6] KVM: lapic: do not set KVM_REQ_EVENT unnecessarily on
 EOI



On 19/12/2016 10:47, Paolo Bonzini wrote:
> On EOI, there is no need to set KVM_REQ_EVENT unconditionally.  The PPR
> update is already setting it if resetting the ISR bit causes PPR to
> decrease.  Even a level-triggered IOAPIC interrupt will set KVM_REQ_EVENT
> on a reinjection (ioapic_service -> kvm_irq_delivery_to_apic and from
> there to __apic_accept_irq).
> 
> Signed-off-by: Paolo Bonzini <pbonzini@...hat.com>
> ---
>  arch/x86/kvm/lapic.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
> index e35cbd44c505..70c7428b7a57 100644
> --- a/arch/x86/kvm/lapic.c
> +++ b/arch/x86/kvm/lapic.c
> @@ -1066,7 +1066,6 @@ static int apic_set_eoi(struct kvm_lapic *apic)
>  		kvm_hv_synic_send_eoi(apic->vcpu, vector);
>  
>  	kvm_ioapic_send_eoi(apic, vector);
> -	kvm_make_request(KVM_REQ_EVENT, apic->vcpu);
>  	return vector;
>  }
>  
> @@ -1081,7 +1080,6 @@ void kvm_apic_set_eoi_accelerated(struct kvm_vcpu *vcpu, int vector)
>  	trace_kvm_eoi(apic, vector);
>  
>  	kvm_ioapic_send_eoi(apic, vector);
> -	kvm_make_request(KVM_REQ_EVENT, apic->vcpu);
>  }
>  EXPORT_SYMBOL_GPL(kvm_apic_set_eoi_accelerated);
>  
> 

This patch breaks Windows XP and 2003 (reverting it while keeping the
others results in 3 successful installs, compared to a failure rate of
12/13 with the patch included).

Paolo

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ