linux-kernel - Re: [PATCH] KVM: x86: Update Xen TSC leaves during CPUID emulation

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: <c7b494d34d3e14531337311c286a9c06a99c9295.camel@infradead.org>
Date: Thu, 23 Jan 2025 12:32:27 -0800
From: David Woodhouse <dwmw2@...radead.org>
To: Fred Griffoul <fgriffo@...zon.co.uk>, kvm@...r.kernel.org
Cc: griffoul@...il.com, vkuznets@...hat.com, Sean Christopherson
 <seanjc@...gle.com>, Paolo Bonzini <pbonzini@...hat.com>, Thomas Gleixner
 <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>, Borislav Petkov
 <bp@...en8.de>, Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
 "H. Peter Anvin" <hpa@...or.com>, Paul Durrant <paul@....org>,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH] KVM: x86: Update Xen TSC leaves during CPUID emulation

On Thu, 2025-01-23 at 19:02 +0000, Fred Griffoul wrote:

> +static inline void kvm_xen_may_update_tsc_info(struct kvm_vcpu *vcpu,
> +					       u32 function, u32 index,
> +					       u32 *eax, u32 *ecx, u32 *edx)

Should this be called kvm_xen_maybe_update_tsc_info() ? 

Is it worth adding if (static_branch_unlikely(&kvm_xen_enabled.key))? 

> +{
> +	u32 base = vcpu->arch.xen.cpuid.base;
> +
> +	if (base && (function == (base | XEN_CPUID_LEAF(3)))) {
> +		if (index == 1) {
> +			*ecx = vcpu->arch.hv_clock.tsc_to_system_mul;
> +			*edx = vcpu->arch.hv_clock.tsc_shift;

Are these fields in vcpu->arch.hv_clock definitely going to be set? If
so, can we have a comment to that effect? And perhaps a warning to
assert the truth of that claim?

Before this patch, if the hv_clock isn't yet set then the guest would
see the original content of the leaves as set by userspace? Now it gets
zeroes if that happens?

Download attachment "smime.p7s" of type "application/pkcs7-signature" (5069 bytes)