Re: [PATCH 5/5] x86/virt/tdx: Export global metadata read infrastructure

From: "Huang, Kai" <kai.huang@intel.com>
To: "kvm@vger.kernel.org" <kvm@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"Edgecombe, Rick P" <rick.p.edgecombe@intel.com>
Cc: "Hansen, Dave" <dave.hansen@intel.com>,
	"seanjc@google.com" <seanjc@google.com>,
	"bp@alien8.de" <bp@alien8.de>, "x86@kernel.org" <x86@kernel.org>,
	"peterz@infradead.org" <peterz@infradead.org>,
	"hpa@zytor.com" <hpa@zytor.com>,
	"mingo@redhat.com" <mingo@redhat.com>,
	"kirill.shutemov@linux.intel.com"
	<kirill.shutemov@linux.intel.com>,
	"tglx@linutronix.de" <tglx@linutronix.de>,
	"pbonzini@redhat.com" <pbonzini@redhat.com>,
	"jgross@suse.com" <jgross@suse.com>,
	"Yamahata, Isaku" <isaku.yamahata@intel.com>
Subject: Re: [PATCH 5/5] x86/virt/tdx: Export global metadata read infrastructure
Date: Mon, 6 May 2024 11:35:10 +0000	[thread overview]
Message-ID: <eacb326a450be89fa49fb90bdfea1f2ed415685a.camel@intel.com> (raw)
In-Reply-To: <6e839de0e66ac2a9a8af7fd78757b2ed1bffbca4.camel@intel.com>

On Fri, 2024-05-03 at 18:31 +0000, Edgecombe, Rick P wrote:
> On Fri, 2024-05-03 at 14:17 +1200, Huang, Kai wrote:
> > 
> > 
> > On 3/05/2024 12:21 pm, Edgecombe, Rick P wrote:
> > > On Sat, 2024-03-02 at 00:20 +1300, Kai Huang wrote:
> > > > KVM will need to read a bunch of non-TDMR related metadata to create and
> > > > run TDX guests.  Export the metadata read infrastructure for KVM to use.
> > > > 
> > > > Specifically, export two helpers:
> > > > 
> > > > 1) The helper which reads multiple metadata fields to a buffer of a
> > > >      structure based on the "field ID -> structure member" mapping table.
> > > > 
> > > > 2) The low level helper which just reads a given field ID.
> > > > 
> > > Could 2 be a static inline in a helper, and then only have one export?
> > 
> > I assume you were thinking about making 2) call the 1), so we don't need 
> > to export 2).
> > 
> > This is not feasible due to:
> > 
> > a). 1) must 'struct tdx_metadata_field_mapping' table as input, and for 
> > that we need to ues the TDX_SYSINFO_MAP() macro to define a structure 
> > for just one 'u64' and define a 'struct tdx_metadata_field_mapping' 
> > table which only has one element for that.
> > 
> > b). TDX_SYSINFO_MAP() macro actually doesn't support taking a metadata 
> > field "variable", but can only support using the name of the metadata field.
> > 
> > However we can try to make 1) as a wrapper of 2).  But this would 
> > require some change to the patch 4.
> > 
> > I'll reply separately to patch 4 and you can take a look whether that is 
> > better.
> 
> Having one export would be nice. Yea, since the multiple field processing
> version just loops anyway, doing it like you propose seems reasonable.

Sure I'll switch to use the new code as replied in patch 4.