Re: [RFC PATCH 05/15] virt: tdx-guest: Add TDCM helpers and TEE-IO support check
From: Nikolay Borisov
Date: Wed Oct 28 2026 - 02:52:01 EST
On 28.10.26 г. 7:42 ч., Duan, Zhenzhong wrote:
-----Original Message-----
From: Nikolay Borisov <nik.borisov@xxxxxxxx>
Subject: Re: [RFC PATCH 05/15] virt: tdx-guest: Add TDCM helpers and TEE-IO
support check
On 26.10.26 г. 3:26 ч., Xiaoyao Li wrote:
On 10/14/2026 2:11 PM, Duan, Zhenzhong wrote:order
This is used to allow TD guest TEE-IO Device Control and Management
(TDCM) module to
manage the device. The TDCM leaf functions are defined in the following
table
Is TDCM module implemented via the NRX facility ? The TDX Connect spec
never mentions the TDCM module. A bit more details are required in
module"to better understand why you've implemented this the way you have."TDCM module" is GHCI terminology. In this series, the guest-side
implementation is in the TDX TSM driver; it is not an NRX module.
I think the TDX guest TSM driver is just the implementation of the "TDCM
called in GHCI, right?
I think we are entering a dangerous bikeshedding territory here. But
IIUC TDCM seems to be something which the TD calls, because it sends the
TDCM function ids to the host, so the TDCM should really be implemented
by the host. But I think such construct just adds more confusion, TDCM
is just a subset of the GHCI, so nothing really more special.
GHCI spec says:
"This is used to allow TD guest TEE-IO Device Control and Management (TDCM) module to
manage the device."
So I think "TDCM module" term in GHCI spec is talking about guest side - the TDX TSM driver.
Since "TDCM module" term is confusing for reader, so I plan to drop it
in next version, I plan to use "TDCM request" to denote the guest-side
operations and "TDCM implementation" to denote the host-side logic.
let me know if this distinction makes sense to you all, or if you have
any other suggestions.
Yep, makes sense to me.