Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v2 1/2] KVM: x86: Add a framework for supporting MSR-based
 features
To:     Tom Lendacky <thomas.lendacky@amd.com>, x86@kernel.org,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org
Cc:     Joerg Roedel <joro@8bytes.org>, Borislav Petkov <bp@alien8.de>,
        Thomas Gleixner <tglx@linutronix.de>,
        =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= <rkrcmar@redhat.com>
References: <20180215231156.31016.79657.stgit@tlendack-t1.amdoffice.net>
 <20180215231206.31016.78914.stgit@tlendack-t1.amdoffice.net>
 <00583b68-d599-b709-133a-3741c258df13@redhat.com>
 <7f6a78ae-abab-2a03-aebd-e43e6fe168b4@amd.com>
 <94f6ed44-07ab-ff0f-2a6d-b3d8ff4ff4b1@redhat.com>
 <f1b934f4-e228-e6e3-496d-d8287018602e@amd.com>
 <d77a8a6e-d10b-82d9-6df4-7ed5505ef40f@amd.com>
From:   Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <ff8260fc-a928-31f8-38ab-248fcc068193@redhat.com>
Date:   Wed, 21 Feb 2018 17:36:27 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.6.0
MIME-Version: 1.0
In-Reply-To: <d77a8a6e-d10b-82d9-6df4-7ed5505ef40f@amd.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 21/02/2018 15:52, Tom Lendacky wrote:
> On 2/21/2018 8:47 AM, Tom Lendacky wrote:
>> On 2/21/2018 8:32 AM, Paolo Bonzini wrote:
>>> On 21/02/2018 15:15, Tom Lendacky wrote:
>>>> On 2/21/2018 5:41 AM, Paolo Bonzini wrote:
>>>>> On 16/02/2018 00:12, Tom Lendacky wrote:
>>>>>> +static u32 msr_based_features[] = {
>>>>>> +};
>>>>>> +
>>>>>> +static unsigned int num_msr_based_features = ARRAY_SIZE(msr_based_features);
>>>>>> +
>>>>>>  bool kvm_valid_efer(struct kvm_vcpu *vcpu, u64 efer)
>>>>>>  {
>>>>>>  	if (efer & efer_reserved_bits)
>>>>>> @@ -2785,6 +2794,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>>>>>>  	case KVM_CAP_SET_BOOT_CPU_ID:
>>>>>>   	case KVM_CAP_SPLIT_IRQCHIP:
>>>>>>  	case KVM_CAP_IMMEDIATE_EXIT:
>>>>>> +	case KVM_CAP_GET_MSR_FEATURES:
>>>>>>  		r = 1;
>>>>>>  		break;
>>>>>>  	case KVM_CAP_ADJUST_CLOCK:
>>>>>> @@ -4410,6 +4420,47 @@ long kvm_arch_vm_ioctl(struct file *filp,
>>>>>>  			r = kvm_x86_ops->mem_enc_unreg_region(kvm, &region);
>>>>>>  		break;
>>>>>>  	}
>>>>>> +	case KVM_GET_MSR_INDEX_LIST: {
>>>>>> +		struct kvm_msr_list __user *user_msr_list = argp;
>>>>>> +		struct kvm_msr_list msr_list;
>>>>>> +		unsigned int n;
>>>>>> +
>>>>>> +		r = -EFAULT;
>>>>>> +		if (copy_from_user(&msr_list, user_msr_list, sizeof(msr_list)))
>>>>>> +			goto out;
>>>>>> +		n = msr_list.nmsrs;
>>>>>> +		msr_list.nmsrs = num_msr_based_features;
>>>>>> +		if (copy_to_user(user_msr_list, &msr_list, sizeof(msr_list)))
>>>>>> +			goto out;
>>>>>> +		r = -E2BIG;
>>>>>> +		if (n < msr_list.nmsrs)
>>>>>> +			goto out;
>>>>>> +		r = -EFAULT;
>>>>>> +		if (copy_to_user(user_msr_list->indices, &msr_based_features,
>>>>>> +				 num_msr_based_features * sizeof(u32)))
>>>>>> +			goto out;
>>>>>> +		r = 0;
>>>>>> +		break;
>>>>>
>>>>> I think it's better to have some logic in kvm_init_msr_list, to filter
>>>>> the MSR list based on whatever MSRs the backend provides.
>>>>
>>>> Ok, that's what I had originally and then you said to just return the full
>>>> list and let KVM_GET_MSR return a 0 or 1 if it was supported. I can switch
>>>> it back.
>>>
>>> Hmm, I cannot find this remark (I would have been very confused, so I
>>> tried to look for it).  I commented on removing kvm_valid_msr_feature,
>>> but not kvm_init_msr_list.
>>
>> I think this is the reply that sent me off on that track:
>>   https://marc.info/?l=linux-kernel&m=151862648123153&w=2

Yeah, it was referring to AMD hosts that don't have the MSR.  Sorry for
the confusion.

>> I'll make it consistent with the other MSR-related items and initialize
>> the list in kvm_init_msr_list().  I'll change the signature of the
>> msr_feature() kvm_x86_ops callback to take an index and optionally return
>> a data value so it can be used to check for support when building the
>> list and return a value when needed.
> 
> Hmm, actually I'll just leave the signature alone and pass in a local
> kvm_msr_entry struct variable for the call when initializing the list.

Sounds good!

Paolo