Linux Confidential Computing Development
 help / color / mirror / Atom feed
From: Chao Gao <chao.gao@intel.com>
To: Dave Hansen <dave.hansen@intel.com>
Cc: <kvm@vger.kernel.org>, <linux-coco@lists.linux.dev>,
	<linux-kernel@vger.kernel.org>, <binbin.wu@linux.intel.com>,
	<dave.hansen@linux.intel.com>, <djbw@kernel.org>,
	<ira.weiny@intel.com>, <kai.huang@intel.com>, <kas@kernel.org>,
	<nik.borisov@suse.com>, <paulmck@kernel.org>,
	<pbonzini@redhat.com>, <reinette.chatre@intel.com>,
	<rick.p.edgecombe@intel.com>, <sagis@google.com>,
	<seanjc@google.com>, <tony.lindgren@linux.intel.com>,
	<vannapurve@google.com>, <vishal.l.verma@intel.com>,
	<yilun.xu@linux.intel.com>, <xiaoyao.li@intel.com>,
	<yan.y.zhao@intel.com>, Thomas Gleixner <tglx@kernel.org>,
	Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
	<x86@kernel.org>, "H. Peter Anvin" <hpa@zytor.com>
Subject: Re: [PATCH v9 01/23] x86/virt/tdx: Consolidate TDX global initialization states
Date: Mon, 18 May 2026 15:43:29 +0800	[thread overview]
Message-ID: <agrDIa/foXa2PhCh@intel.com> (raw)
In-Reply-To: <ab35e6e4-b57a-4d49-8521-ca69a9135d3c@intel.com>

On Fri, May 15, 2026 at 09:14:02AM -0700, Dave Hansen wrote:
>On 5/13/26 08:09, Chao Gao wrote:
>...
>> Group the states into a single structure so they can be reset together, for
>> example with memset(), and so a newly added state won't be missed.
>...
>> +struct tdx_module_state {
>> +	bool initialized;
>> +	bool sysinit_done;
>> +	int sysinit_ret;
>> +};
>...
>> @@ -113,30 +119,28 @@ static int try_init_module_global(void)
>>  {
>>  	struct tdx_module_args args = {};
>>  	static DEFINE_RAW_SPINLOCK(sysinit_lock);
>> -	static bool sysinit_done;
>> -	static int sysinit_ret;
>
>This doesn't look right to me.
>
>>  	raw_spin_lock(&sysinit_lock);
>>  
>> -	if (sysinit_done)
>> +	if (tdx_module_state.sysinit_done)
>>  		goto out;
>>  
>>  	/* RCX is module attributes and all bits are reserved */
>>  	args.rcx = 0;
>> -	sysinit_ret = seamcall_prerr(TDH_SYS_INIT, &args);
>> +	tdx_module_state.sysinit_ret = seamcall_prerr(TDH_SYS_INIT, &args);
>>  
>>  	/*
>>  	 * The first SEAMCALL also detects the TDX module, thus
>>  	 * it can fail due to the TDX module is not loaded.
>>  	 * Dump message to let the user know.
>>  	 */
>> -	if (sysinit_ret == -ENODEV)
>> +	if (tdx_module_state.sysinit_ret == -ENODEV)
>>  		pr_err("module not loaded\n");
>>  
>> -	sysinit_done = true;
>> +	tdx_module_state.sysinit_done = true;
>>  out:
>>  	raw_spin_unlock(&sysinit_lock);
>> -	return sysinit_ret;
>> +	return tdx_module_state.sysinit_ret;
>>  }
>
>But I think it's because 'sysinit_ret' is really a funky thing. It's
>just here so that the module only _tries_ to be initialized once. That
>one-time init records its error code in sysinit_ret and then secondary
>callers pick it up.
>
>Here's how you do that in a non-confusing way (some context chopped out):
>
>static int try_init_module_global(void)
>{
>        struct tdx_module_args args = {};
>        static DEFINE_RAW_SPINLOCK(sysinit_lock);
>        int ret;
>
>        raw_spin_lock(&sysinit_lock);
>
>	/* Return the "cached" return code: */
>        if (tdx_module_state.sysinit_done) {
>		ret = tdx_module_state.sysinit_ret;
>                goto out;
>	}
>
>        ret = seamcall_prerr(TDH_SYS_INIT, &args);
>
>	/* Save the return code for later callers: */
>	tdx_module_state.sysinit_ret  = ret;
>        tdx_module_state.sysinit_done = true;
>out:
>        raw_spin_unlock(&sysinit_lock);
>        return ret;
>}
>
>See how it sets the module state in _one_ place? It also only touches
>the module state under the lock so it's more obvious that it is correct
>and there are no races or tearing or other nonsense.
>
>*That* is a proper refactoring.

Yes, that is a clear readability improvement.

To follow the "one change per patch" rule, I am inclined to do that as a
separate preparatory cleanup before moving all states into a structure.

commit 26bb389c5762fd6a496fbed1cc55e4978e99a5cb
Author: Chao Gao <chao.gao@intel.com>
Date:   Sun May 17 20:03:00 2026 -0700

    x86/virt/tdx: Clarify try_init_module_global() result caching
    
    TDX module global initialization is executed only once. The first call
    caches both the result and the "done" state, and later callers reuse the
    saved result. A lock protects that cached state.
    
    The current code is harder to read because sysinit_done is accessed under
    the lock, while sysinit_ret is not.
    
    To improve readability, move sysinit_ret accesses within the lock.
    
    Group sysinit_ret/sysinit_done updates right after initialization so
    Caching the result is separate from the initialization itself.
    
    Signed-off-by: Chao Gao <chao.gao@intel.com>

diff --git a/arch/x86/virt/vmx/tdx/tdx.c b/arch/x86/virt/vmx/tdx/tdx.c
index c0c6281b08a5..ad56f142dd0b 100644
--- a/arch/x86/virt/vmx/tdx/tdx.c
+++ b/arch/x86/virt/vmx/tdx/tdx.c
@@ -115,28 +115,34 @@ static int try_init_module_global(void)
	static DEFINE_RAW_SPINLOCK(sysinit_lock);
	static bool sysinit_done;
	static int sysinit_ret;
+	int ret;
 
	raw_spin_lock(&sysinit_lock);
 
-	if (sysinit_done)
+	/* Return the "cached" return code. */
+	if (sysinit_done) {
+		ret = sysinit_ret;
		goto out;
+	}
 
	/* RCX is module attributes and all bits are reserved */
	args.rcx = 0;
-	sysinit_ret = seamcall_prerr(TDH_SYS_INIT, &args);
+	ret = seamcall_prerr(TDH_SYS_INIT, &args);
 
	/*
	 * The first SEAMCALL also detects the TDX module, thus
	 * it can fail due to the TDX module is not loaded.
	 * Dump message to let the user know.
	 */
-	if (sysinit_ret == -ENODEV)
+	if (ret == -ENODEV)
		pr_err("module not loaded\n");
 
+	/* Save the return code for later callers. */
	sysinit_done = true;
+	sysinit_ret = ret;
 out:
	raw_spin_unlock(&sysinit_lock);
-	return sysinit_ret;
+	return ret;
 }
 
 /**

  reply	other threads:[~2026-05-18  7:43 UTC|newest]

Thread overview: 66+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-13 15:09 [PATCH v9 00/23] Runtime TDX module update support Chao Gao
2026-05-13 15:09 ` [PATCH v9 01/23] x86/virt/tdx: Consolidate TDX global initialization states Chao Gao
2026-05-15 16:14   ` Dave Hansen
2026-05-18  7:43     ` Chao Gao [this message]
2026-05-18 18:00       ` Edgecombe, Rick P
2026-05-18 18:09         ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 02/23] x86/virt/tdx: Move TDX_FEATURES0 bits to asm/tdx.h Chao Gao
2026-05-15 16:15   ` Dave Hansen
2026-05-18  7:52     ` Chao Gao
2026-05-18 16:57       ` Edgecombe, Rick P
2026-05-19  1:59         ` Edgecombe, Rick P
2026-05-19 10:24           ` Chao Gao
2026-05-13 15:09 ` [PATCH v9 03/23] x86/virt/tdx: Move low level SEAMCALL helpers out of <asm/tdx.h> Chao Gao
2026-05-13 15:09 ` [PATCH v9 04/23] coco/tdx-host: Introduce a "tdx_host" device Chao Gao
2026-05-15 16:21   ` Dave Hansen
2026-05-18 11:18     ` Chao Gao
2026-05-18 18:08       ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 05/23] coco/tdx-host: Expose TDX module version Chao Gao
2026-05-15 16:53   ` Dave Hansen
2026-05-18 11:29     ` Chao Gao
2026-05-13 15:09 ` [PATCH v9 06/23] x86/virt/seamldr: Introduce a wrapper for P-SEAMLDR SEAMCALLs Chao Gao
2026-05-15 17:02   ` Dave Hansen
2026-05-15 17:07   ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 07/23] x86/virt/seamldr: Add a helper to retrieve P-SEAMLDR information Chao Gao
2026-05-15 17:18   ` Dave Hansen
2026-05-18 11:51     ` Chao Gao
2026-05-13 15:09 ` [PATCH v9 08/23] coco/tdx-host: Expose P-SEAMLDR information via sysfs Chao Gao
2026-05-15 17:23   ` Dave Hansen
2026-05-18 12:35     ` Chao Gao
2026-05-15 17:28   ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 09/23] coco/tdx-host: Don't expose P-SEAMLDR information on CPUs with erratum Chao Gao
2026-05-15 17:26   ` Dave Hansen
2026-05-18 12:44     ` Chao Gao
2026-05-18 15:29       ` Dave Hansen
2026-05-19  1:22         ` Edgecombe, Rick P
2026-05-13 15:09 ` [PATCH v9 10/23] coco/tdx-host: Implement firmware upload sysfs ABI for TDX module updates Chao Gao
2026-05-15 18:05   ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 11/23] x86/virt/seamldr: Allocate and populate a module update request Chao Gao
2026-05-15  6:05   ` Chao Gao
2026-05-15 18:24   ` Dave Hansen
2026-05-15 18:44     ` Edgecombe, Rick P
2026-05-15 18:59       ` Dave Hansen
2026-05-18 14:15     ` Chao Gao
2026-05-18 15:12       ` Dave Hansen
2026-05-15 19:43   ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 12/23] x86/virt/seamldr: Introduce skeleton for TDX module updates Chao Gao
2026-05-18 15:36   ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 13/23] x86/virt/seamldr: Abort updates after a failed step Chao Gao
2026-05-19  2:34   ` Edgecombe, Rick P
2026-05-19 10:20     ` Chao Gao
2026-05-13 15:09 ` [PATCH v9 14/23] x86/virt/seamldr: Shut down the current TDX module Chao Gao
2026-05-19  3:00   ` Edgecombe, Rick P
2026-05-19 12:05     ` Chao Gao
2026-05-19 16:24       ` Dave Hansen
2026-05-13 15:09 ` [PATCH v9 15/23] x86/virt/tdx: Reset software states during TDX module shutdown Chao Gao
2026-05-13 15:09 ` [PATCH v9 16/23] x86/virt/seamldr: Install a new TDX module Chao Gao
2026-05-13 15:10 ` [PATCH v9 17/23] x86/virt/seamldr: Do TDX per-CPU initialization after module installation Chao Gao
2026-05-13 15:10 ` [PATCH v9 18/23] x86/virt/tdx: Restore TDX module state Chao Gao
2026-05-13 15:10 ` [PATCH v9 19/23] x86/virt/tdx: Refresh TDX module version after update Chao Gao
2026-05-19  3:16   ` Edgecombe, Rick P
2026-05-19 10:42     ` Chao Gao
2026-05-13 15:10 ` [PATCH v9 20/23] x86/virt/tdx: Reject updates during compatibility-sensitive operations Chao Gao
2026-05-15  6:12   ` Chao Gao
2026-05-13 15:10 ` [PATCH v9 21/23] x86/virt/tdx: Enable TDX module runtime updates Chao Gao
2026-05-13 15:10 ` [PATCH v9 22/23] coco/tdx-host: Document TDX module update compatibility criteria Chao Gao
2026-05-13 15:10 ` [PATCH v9 23/23] x86/virt/tdx: Document TDX module update Chao Gao

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=agrDIa/foXa2PhCh@intel.com \
    --to=chao.gao@intel.com \
    --cc=binbin.wu@linux.intel.com \
    --cc=bp@alien8.de \
    --cc=dave.hansen@intel.com \
    --cc=dave.hansen@linux.intel.com \
    --cc=djbw@kernel.org \
    --cc=hpa@zytor.com \
    --cc=ira.weiny@intel.com \
    --cc=kai.huang@intel.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=nik.borisov@suse.com \
    --cc=paulmck@kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=reinette.chatre@intel.com \
    --cc=rick.p.edgecombe@intel.com \
    --cc=sagis@google.com \
    --cc=seanjc@google.com \
    --cc=tglx@kernel.org \
    --cc=tony.lindgren@linux.intel.com \
    --cc=vannapurve@google.com \
    --cc=vishal.l.verma@intel.com \
    --cc=x86@kernel.org \
    --cc=xiaoyao.li@intel.com \
    --cc=yan.y.zhao@intel.com \
    --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