linux-kernel - Re: x86/smpboot: Question regarding native_play_dead() _

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4be37a6a-0a04-48f8-9399-25ecb4638cdc@redhat.com>
Date: Mon, 27 Oct 2025 11:48:28 -0400
From: Waiman Long <llong@...hat.com>
To: Peter Zijlstra <peterz@...radead.org>,
 Thorsten Blum <thorsten.blum@...ux.dev>
Cc: Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>,
 Borislav Petkov <bp@...en8.de>, Dave Hansen <dave.hansen@...ux.intel.com>,
 "H. Peter Anvin" <hpa@...or.com>, linux-kernel@...r.kernel.org,
 x86@...nel.org
Subject: Re: x86/smpboot: Question regarding native_play_dead() __noreturn
 warning

On 10/27/25 8:50 AM, Peter Zijlstra wrote:
> On Mon, Oct 27, 2025 at 01:23:02PM +0100, Thorsten Blum wrote:
>> Hi,
>>
>> I just came across this comment in arch/x86/kernel/smpboot.c:
>>
>> /*
>> * native_play_dead() is essentially a __noreturn function, but it can't
>> * be marked as such as the compiler may complain about it.
>> */
>> void native_play_dead(void) {
>> 	...
>> }
>>
>> and when I mark native_play_dead() as __noreturn, neither gcc nor clang
>> complain about it.
>>
>> The commit message 2743fe89d4d4 ("x86/idle: Disable IBRS when CPU is
>> offline to improve single-threaded performance") says:
>>
>> "Add a comment to say that native_play_dead() is a __noreturn function,
>> but it can't be marked as such to avoid confusion about the missing
>> MSR restoration code."
>>
>> Unfortunately, that doesn't really help me either. Can someone explain
>> what the issue was and if the comment is still valid? Otherwise, I'd
>> like to submit a patch adding __noreturn and removing the comment.
> I'm not sure either, it wasn't there in v2 but appeared in v3.
>
> v2: 20230620140625.1001886-3-longman@...hat.com
> v3: 20230622003603.1188364-2-longman@...hat.com
>
> The difference is that v2 tried to restore the msr after 'play_dead'
> which is silly, since it would never reach that code. v3 removed that
> dead restore code and added the confusing comment.
>
> There is a clue here though:
>
>    20230622054053.uy577qezu5a65buc@...ble
>
> Josh suggests play_dead() should be marked noreturn (which it is in
> current kernels).
>
> Waiman then replies:
>
>    921e1b98-af36-1f51-5abe-dea36425b706@...hat.com
>
> which is utterly confused again.

I don't remember exactly how I got the warning when __noreturn is added 
to native_play_dead(). It may be a limitation of the objtool or gcc that 
I was using at that time. If Thorsten doesn't have problem adding 
__noreturn, I won't mind him doing that and taking out the the comment. 
We can see if there is other issue coming up in the future.

Cheers,
Longman