linux-kernel - Re: [PATCH v7 13/15] KVM: x86/vmx: Clear Arch LBREn bit before inject #DB to guest

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20210809090215.GA31980@intel.com>
Date:   Mon, 9 Aug 2021 17:02:15 +0800
From:   Yang Weijiang <weijiang.yang@...el.com>
To:     Like Xu <like.xu.linux@...il.com>
Cc:     Yang Weijiang <weijiang.yang@...el.com>, pbonzini@...hat.com,
        jmattson@...gle.com, seanjc@...gle.com, vkuznets@...hat.com,
        wei.w.wang@...el.com, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v7 13/15] KVM: x86/vmx: Clear Arch LBREn bit before
 inject #DB to guest

On Mon, Aug 09, 2021 at 01:08:32PM +0800, Like Xu wrote:
> On 6/8/2021 3:42 pm, Yang Weijiang wrote:
> >Per ISA spec, need to clear the bit before inject #DB.
> 
> Please paste the SDM statement accurately so that the reviewers
> can verify that the code is consistent with the documentation.
>
Thanks Like! Sure, will add the description in commit message.

> >
> >Signed-off-by: Yang Weijiang <weijiang.yang@...el.com>
> >---
> >  arch/x86/kvm/vmx/vmx.c | 21 +++++++++++++++++++++
> >  1 file changed, 21 insertions(+)
> >
> >diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> >index 70314cd93340..31b9c06c9b3b 100644
> >--- a/arch/x86/kvm/vmx/vmx.c
> >+++ b/arch/x86/kvm/vmx/vmx.c
> >@@ -1601,6 +1601,21 @@ static void vmx_clear_hlt(struct kvm_vcpu *vcpu)
> >  		vmcs_write32(GUEST_ACTIVITY_STATE, GUEST_ACTIVITY_ACTIVE);
> >  }
> >+static void flip_arch_lbr_ctl(struct kvm_vcpu *vcpu, bool on)
> >+{
> >+	if (vcpu_to_pmu(vcpu)->event_count > 0 &&
> 
> Ugh, this check seems ridiculous/funny to me.
Do you expect aditional bit-check for INTEL_PMC_IDX_FIXED_VLBR in 
pmu->pmc_in_use?

> 
> >+	    kvm_cpu_cap_has(X86_FEATURE_ARCH_LBR)) {
> >+		u64 lbr_ctl = vmcs_read64(GUEST_IA32_LBR_CTL);
> >+
> >+		if (on)
> >+			lbr_ctl |= 1ULL;
> >+		else
> >+			lbr_ctl &= ~1ULL;
> >+
> >+		vmcs_write64(GUEST_IA32_LBR_CTL, lbr_ctl);
> >+	}
> >+}
> 
> ...