Subject: Re: [PATCH 1/7] KVM: x86: add SPEC_CTRL and IBPB_SUPPORT accessors
To: Liran Alon <LIRAN.ALON@ORACLE.COM>, linux-kernel@vger.kernel.org,
        kvm@vger.kernel.org
Cc: jmattson@google.com, aliguori@amazon.com, thomas.lendacky@amd.com,
        dwmw@amazon.co.uk, bp@alien8.de
References: <1515434925-10250-1-git-send-email-pbonzini@redhat.com>
 <1515434925-10250-2-git-send-email-pbonzini@redhat.com>
 <5A53C1DA.50901@ORACLE.COM>
From: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <91f7f729-1f45-8081-4ac6-af0c82b712e4@redhat.com>
Date: Tue, 9 Jan 2018 11:32:31 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.5.0
MIME-Version: 1.0
In-Reply-To: <5A53C1DA.50901@ORACLE.COM>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org

On 08/01/2018 20:09, Liran Alon wrote:
>>
>> +static inline bool cpu_has_spec_ctrl(void)
>> +{
>> +    u32 eax, ebx, ecx, edx;
>> +    cpuid_count(7, 0, &eax, &ebx, &ecx, &edx);
>> +
>> +    return edx & bit(KVM_CPUID_BIT_SPEC_CTRL);
> 
> Why not just "return cpuid_edx(7) & bit(KVM_CPUID_BIT_SPEC_CTRL);"?
> This is also consistent with how cpu_has_ibpb_support() is written.

Leaf 7 explicitly requires you to clear ECX (there could be a leaf for
EAX=7,ECX=1 in the future).  Even though cpuid_edx does do that, it's
not clear from the function that cpuid_edx(7) would work.

Paolo