linux-kernel - Re: [PATCH v1 06/11] KVM: x86: use hardware-compatible format for APIC ID register

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: <217814b5-d637-07f8-48cd-777ad498c54f@redhat.com>
Date:	Fri, 1 Jul 2016 16:12:09 +0200
From:	Paolo Bonzini <pbonzini@...hat.com>
To:	Radim Krčmář <rkrcmar@...hat.com>
Cc:	linux-kernel@...r.kernel.org, kvm@...r.kernel.org,
	"Lan, Tianyu" <tianyu.lan@...el.com>,
	Igor Mammedov <imammedo@...hat.com>,
	Jan Kiszka <jan.kiszka@....de>, Peter Xu <peterx@...hat.com>
Subject: Re: [PATCH v1 06/11] KVM: x86: use hardware-compatible format for
 APIC ID register



On 01/07/2016 15:11, Radim Krčmář wrote:
>>> >> +static void __kvm_apic_state_fixup(struct kvm_vcpu *vcpu,
>>> >> +		struct kvm_lapic_state *s, bool set)
>>> >> +{
>>> >> +	if (apic_x2apic_mode(vcpu->arch.apic)) {
>>> >> +		u32 *id = (u32 *)(s->regs + APIC_ID);
>>> >> +		if (set)
>>> >> +			*id >>= 24;
>>> >> +		else
>>> >> +			*id <<= 24;
>>> >> +	}
>> > 
>> > When setting, this should read from the apic_base being set.  So I think
>> > you cannot use apic_x2apic_mode.
> 
> apic_x2apic_mode uses apic_base MSR, so its value does not depend on
> LAPIC_SET/GET.  I don't like the dependency much, but all combinations
> of values/orders should work well.

You're right of course.  However it should be documented in the
KVM_GET_LAPIC/KVM_SET_LAPIC ioctl docs that KVM_SET_MSR for apic_base
should be performed first.

Should kvm_lapic_set_base change the value of the ID register when
x2apic mode is left just like we do for entering x2apic mode?  (Hint: we
want kvm-unit-tests for this).

Paolo