Re: [RFC PATCH 02/15] x86/tdx: Add TDCM hypercall wrapper for TDX Connect

[Xiaoyao Li]

On 10/27/2026 2:08 PM, Duan, Zhenzhong wrote:
> 
> 
>> -----Original Message-----
>> From: Li, Xiaoyao <xiaoyao.li@xxxxxxxxx>
>> Subject: Re: [RFC PATCH 02/15] x86/tdx: Add TDCM hypercall wrapper for TDX
>> Connect
>>
>> On 10/27/2026 11:28 AM, Duan, Zhenzhong wrote:
>>>>> Abstract the underlying leaf call and pass the shared buffer physical
>>>>> address as a decrypted mapping. Add a runtime page alignment check
>>>>> for the buffer to enforce correct usage.
>>>> This doesn't match the implementation. The code only WARNs instead of
>>>> enforcing
>>>> correct usage.
>>> You mean below code?
>>>
>>> static void tdx_devsec_unlock(struct pci_tsm *tsm)
>>> {
>>>         struct tdx_devsec *tdevsec = to_tdx_devsec(tsm);
>>>         struct device_evidence *evidence = tsm->evidence;
>>>
>>>         if (WARN_ON(tdx_tdi_unbind_dev(tdevsec)))
>>>                 return;
>>>
>>> I think the unlock() callback should be defined with a return value so
>>> userspace could catch the error.
>>
>> No. I meant just the code of tdx_hcall_tdcm()
>>
>> +u64 tdx_hcall_tdcm(u16 devid, void *buf, size_t size, u8 vector)
>> +{
>> +	WARN_ON_ONCE(!PAGE_ALIGNED(buf) || !PAGE_ALIGNED(size));
>> +
>> +	return _tdx_hypercall(TDVMCALL_TDCM, devid,
>> cc_mkdec(virt_to_phys(buf)), size,
>> vector);
>> +}
>>
>> it warns on the unaligned buf and size, but the unaligned data are still
>> consumed.
> 
> The caller has guaranteed gpa and size are 4KB aligned. 

So far, no caller yet.

> Meanwhile GHCI spec
> requires passing 4KB aligned gpa and size, that means if we don't, the
> hypercall will return failure. So maybe we can drop it?

yeah. It should be OK to drop the WARN_ON_ONCE().