From: "Edgecombe, Rick P" <rick.p.edgecombe@intel.com>
To: "kvm@vger.kernel.org" <kvm@vger.kernel.org>,
"linux-coco@lists.linux.dev" <linux-coco@lists.linux.dev>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"Gao, Chao" <chao.gao@intel.com>
Cc: "bp@alien8.de" <bp@alien8.de>, "x86@kernel.org" <x86@kernel.org>,
"kas@kernel.org" <kas@kernel.org>,
"hpa@zytor.com" <hpa@zytor.com>,
"yilun.xu@linux.intel.com" <yilun.xu@linux.intel.com>,
"mingo@redhat.com" <mingo@redhat.com>,
"dave.hansen@linux.intel.com" <dave.hansen@linux.intel.com>,
"tglx@kernel.org" <tglx@kernel.org>
Subject: Re: [RFC PATCH v2 01/10] x86/virt/tdx: Add a helper to read a table of metadata fields
Date: Wed, 23 Sep 2026 22:20:09 +0000 [thread overview]
Message-ID: <073407d07707e1fd083f238674560c5b7eb1c4c6.camel@intel.com> (raw)
In-Reply-To: <20260918132946.76533-2-chao.gao@intel.com>
On Fri, 2026-09-18 at 06:29 -0700, Chao Gao wrote:
> The metadata field readers get_tdx_sys_info_<class>() in
> tdx_global_metadata.c were generated by an out-of-tree script. That has
> not worked out: the JSON file they were generated from is neither stable
> nor authoritative enough [1]. The goal now is to maintain the readers by
> hand and to establish one standard way of adding a metadata field.
>
> Take get_tdx_sys_info_version() as an example:
>
> if (!ret && !(ret = read_sys_metadata_field(0x0800000100000003, &val)))
> sysinfo_version->minor_version = val;
> if (!ret && !(ret = read_sys_metadata_field(0x0800000100000004, &val)))
> sysinfo_version->major_version = val;
> if (!ret && !(ret = read_sys_metadata_field(0x0800000100000005, &val)))
> sysinfo_version->update_version = val;
>
> Two patterns stand out: the read-check-store sequence repeats once per
> field, and the error of each read is chained into the reads that follow.
> Neither is common in hand-written code.
>
> Eliminate both with a loop that reads each field, stores the value into
> its structure member, and returns on the first error.
"Prepare to eliminate both..."?
>
> Add 'struct field_mapping' to describe one field as its ID plus the
> offset and size of the member that receives its value. Add
> TDX_SYSINFO_MAP() to build such an entry from a field ID name, a
> structure type and a member name. Add __read_sys_metadata_table() to
> read every field in a table. Annotate the helper __maybe_unused as there
> is no caller right now.
>
> Following changes will convert the existing readers to use the new helper.
>
> AI was used under supervision to review code and workshop logs.
>
> Signed-off-by: Chao Gao <chao.gao@intel.com>
> Link: https://lore.kernel.org/kvm/1e7bcbad-eb26-44b7-97ca-88ab53467212@intel.com/ # [1]
> ---
> arch/x86/virt/vmx/tdx/tdx.c | 36 ++++++++++++++++++++++++++++++++++++
> 1 file changed, 36 insertions(+)
>
> diff --git a/arch/x86/virt/vmx/tdx/tdx.c b/arch/x86/virt/vmx/tdx/tdx.c
> index 063574ed8625..5cd9c6eb98c2 100644
> --- a/arch/x86/virt/vmx/tdx/tdx.c
> +++ b/arch/x86/virt/vmx/tdx/tdx.c
> @@ -395,6 +395,42 @@ static int read_sys_metadata_field(u64 field_id, u64 *data)
> return 0;
> }
>
> +/*
> + * Map a TDX global metadata field to a structure member.
> + * @field_id: The TDX global metadata field ID.
> + * @size: The size of the structure member.
> + * @offset: The member's offset within its containing structure.
> + */
> +struct field_mapping {
> + u64 field_id;
> + size_t size;
> + int offset;
> +};
> +
> +/* Read each metadata field listed in @mappings[] into @data. */
> +static int __maybe_unused __read_sys_metadata_table(const struct field_mapping *mappings,
> + int num_mappings, void *data)
> +{
> + int i, ret;
> + u64 val;
> +
> + for (i = 0; i < num_mappings; i++) {
> + ret = read_sys_metadata_field(mappings[i].field_id, &val);
> + if (ret)
> + return ret;
> + memcpy((char *)data + mappings[i].offset, &val, mappings[i].size);
> + }
> +
> + return 0;
> +}
> +
> +#define TDX_SYSINFO_MAP(_field, _type, _member) \
> +{ \
> + .field_id = TDX_MD_FIELD_ID_##_field, \
I find it a bit hard to search the code when it constructs the
"TDX_MD_FIELD_ID_" part of the define inside the macro. And since they are all
named "TDX_MD_FIELD_ID_FOO" it seems a bit verbose and repetitive. Can we have a
smaller pattern such that we can use the exact macro name in TDX_SYSINFO_MAP()?
Maybe like this:
#define TDX_FIELD_MODULE_HV 0x8900000100000000ULL
static const struct field_mapping handoff_mappings[] = {
TDX_SYSINFO_MAP_HANDOFF(TDX_FIELD_MODULE_HV, module_hv),
};
> + .offset = offsetof(_type, _member), \
> + .size = sizeof_field(_type, _member), \
> +}
> +
> #include "tdx_global_metadata.c"
>
> static __init int check_features(struct tdx_sys_info *sysinfo)
next prev parent reply other threads:[~2026-09-23 22:20 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-18 13:29 [RFC PATCH v2 00/10] TDX: Stop auto-generating the global metadata code Chao Gao
2026-09-18 13:29 ` [RFC PATCH v2 01/10] x86/virt/tdx: Add a helper to read a table of metadata fields Chao Gao
2026-09-23 22:20 ` Edgecombe, Rick P [this message]
2026-09-18 13:29 ` [RFC PATCH v2 02/10] x86/virt/tdx: Convert the version metadata reader Chao Gao
2026-09-23 22:26 ` Edgecombe, Rick P
2026-09-18 13:29 ` [RFC PATCH v2 03/10] x86/virt/tdx: Convert the features " Chao Gao
2026-09-18 13:29 ` [RFC PATCH v2 04/10] x86/virt/tdx: Convert the tdmr " Chao Gao
2026-09-18 13:29 ` [RFC PATCH v2 05/10] x86/virt/tdx: Convert the td_ctrl " Chao Gao
2026-09-18 13:29 ` [RFC PATCH v2 06/10] x86/virt/tdx: Convert the handoff " Chao Gao
2026-09-18 13:29 ` [RFC PATCH v2 07/10] x86/virt/tdx: Convert the td_conf " Chao Gao
2026-09-18 13:29 ` [RFC PATCH v2 08/10] x86/virt/tdx: Remove tdx_global_metadata.c Chao Gao
2026-09-23 22:53 ` Edgecombe, Rick P
2026-09-18 13:29 ` [RFC PATCH v2 09/10] x86/virt/tdx: Use early returns in get_tdx_sys_info() Chao Gao
2026-09-23 22:52 ` Edgecombe, Rick P
2026-09-18 13:29 ` [RFC PATCH v2 10/10] x86/virt/tdx: Verify structure member sizes against metadata field IDs Chao Gao
2026-09-23 22:47 ` Edgecombe, Rick P
2026-09-23 0:47 ` [RFC PATCH v2 00/10] TDX: Stop auto-generating the global metadata code Edgecombe, Rick P
2026-09-23 8:23 ` Chao Gao
2026-09-23 5:44 ` Tony Lindgren
2026-09-23 8:19 ` Chao Gao
2026-09-23 8:37 ` Tony Lindgren
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=073407d07707e1fd083f238674560c5b7eb1c4c6.camel@intel.com \
--to=rick.p.edgecombe@intel.com \
--cc=bp@alien8.de \
--cc=chao.gao@intel.com \
--cc=dave.hansen@linux.intel.com \
--cc=hpa@zytor.com \
--cc=kas@kernel.org \
--cc=kvm@vger.kernel.org \
--cc=linux-coco@lists.linux.dev \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=tglx@kernel.org \
--cc=x86@kernel.org \
--cc=yilun.xu@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®