Re: [RFC PATCH 02/20] x86/virt/tdx: Prepare to support P-SEAMLDR SEAMCALLs

[Chao Gao <chao.gao@intel.com>]

>Given there will be a dedicated seamldr.c, I don't quite like having
>seamldr_prerr() in "tdx.h" and tdx.c.
>
>Now for all SEAMCALLs used by KVM, we have a dedicated wrapper implemented
>in tdx.c and exported for KVM to use.  I think we can move seamcall*() out
>of <asm/tdx.h> to TDX host local since no other kernel code except the TDX
>host core is supposed to use seamcall*().
>
>This also cleans up <asm/tdx.h> a little bit, which in general makes code
>cleaner IMHO.
>
>E.g., how about we do below patch, and then you can do changes to support
>P-SEAMLDR on top of it?

looks good to me. I'd like to incorporate this patch into my series if
Kirill and Dave have no objections to this cleanup. I assume
seamldr_prerr() can be added to the new seamcall.h

Thanks for this suggestion.

>diff --git a/arch/x86/virt/vmx/tdx/seamcall.h b/arch/x86/virt/vmx/tdx/seamcall.h
>new file mode 100644
>index 000000000000..54922f7bda3a
>--- /dev/null
>+++ b/arch/x86/virt/vmx/tdx/seamcall.h
>@@ -0,0 +1,71 @@
>+/* SPDX-License-Identifier: GPL-2.0 */
>+/* Copyright (C) 2025 Intel Corporation */
>+#include <asm/tdx.h>

If seamcall.h is intended to provide low-level helpers, including
<asm/tdx.h>, which is meant to offer high-level APIs for other components
such as KVM, seems a bit odd to me. But I suppose we can live with this.