Received-SPF: pass (google.com: domain of linux-kernel+bounces-104072-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Message-ID: <0f20ffa7-63ac-4af2-8293-447dcae45eb7@intel.com>
Date: Fri, 15 Mar 2024 13:50:38 +0800
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v19 034/130] KVM: TDX: Get system-wide info about TDX
 module on initialization
Content-Language: en-US
To: "Huang, Kai" <kai.huang@intel.com>,
 "kvm@vger.kernel.org" <kvm@vger.kernel.org>,
 "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
 "Yamahata, Isaku" <isaku.yamahata@intel.com>
Cc: "Zhang, Tina" <tina.zhang@intel.com>, "Yuan, Hang" <hang.yuan@intel.com>,
 "seanjc@google.com" <seanjc@google.com>, "Chen, Bo2" <chen.bo@intel.com>,
 "sagis@google.com" <sagis@google.com>,
 "isaku.yamahata@gmail.com" <isaku.yamahata@gmail.com>,
 "Aktas, Erdem" <erdemaktas@google.com>,
 "pbonzini@redhat.com" <pbonzini@redhat.com>
References: <cover.1708933498.git.isaku.yamahata@intel.com>
 <eaa2c1e23971f058e5921681b0b84d7ea7d38dc1.1708933498.git.isaku.yamahata@intel.com>
 <e88e5448-e354-4ec6-b7de-93dd8f7786b5@intel.com>
 <15a13c5d-df88-46cf-8d88-2c8b94ff41ff@intel.com>
 <aa1ed4c118177e3e341eebccecac3b07bf75a47d.camel@intel.com>
 <10c41a88-d692-4ff5-a0c3-ae13a06a055c@intel.com>
 <078ed2b1c3681c6c0ada9106d481d2f7d964815a.camel@intel.com>
From: Xiaoyao Li <xiaoyao.li@intel.com>
In-Reply-To: <078ed2b1c3681c6c0ada9106d481d2f7d964815a.camel@intel.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

On 3/15/2024 1:39 PM, Huang, Kai wrote:
> On Fri, 2024-03-15 at 13:11 +0800, Li, Xiaoyao wrote:
>>> Here is what Isaku can do using the current API:
>>>
>>>   	u64 num_cpuid_config;
>>   >
>>>
>>>   	...
>>>
>>>   	tdx_sys_metadata_field_read(NUM_CPUID_CONFIG, &num_cpuid_config);
>>>
>>>   	tdx_info = kzalloc(calculate_tdx_info_size(num_cpuid_config), ...);
>>>
>>>   	tdx_info->num_cpuid_config = num_cpuid_config;
>>
>> Dosen't num_cpuid_config serve as temporary variable in some sense?
> 
> You need it, regardless whether it is u64 or u16.
> 
>>
>> For this case, it needs to be used for calculating the size of tdx_info.
>> So we have to have it. But it's not the common case.
>>
>> E.g., if we have another non-u64 field (e.g., field_x) in tdx_info, we
>> cannot to read it via
>>
>> 	tdx_sys_metadata_field_read(FIELD_X_ID, &tdx_info->field_x);
>>
>> we have to use a temporary u64 variable.
> 
> Let me repeat below in my _previous_ reply:
> 
> "
> One example that the current tdx_sys_metadata_field_read() doesn't quite fit is
> you have something like this:
> 
> 	struct {
> 		u16 whatever;
> 		...
> 	} st;
> 
> 	tdx_sys_metadata_field_read(FIELD_ID_WHATEVER, &st.whatever);
> 
> But for this use case you are not supposed to use tdx_sys_metadata_field_read(),
> but use tdx_sys_metadata_read() which has a mapping provided anyway.
> "

tdx_sys_metadata_read() is too complicated for just reading one field.

Caller needs to prepare a one-item size array of "struct 
tdx_metadata_field_mapping" and pass the correct offset.

> So sorry I am not seeing a real example from you.
>