linux-kernel - Re: [PATCH 5.10] x86/retpoline: Move a NOENDBR annotation to the SRSO dummy return thunk

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: <20240708170541.GGZowcZZ0tZmt7znYA@fat_crate.local>
Date: Mon, 8 Jul 2024 19:05:41 +0200
From: Borislav Petkov <bp@...en8.de>
To: Jim Mattson <jmattson@...gle.com>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: Ingo Molnar <mingo@...nel.org>, x86@...nel.org,
	linux-kernel@...r.kernel.org, Greg Thelen <gthelen@...gle.com>
Subject: Re: [PATCH 5.10] x86/retpoline: Move a NOENDBR annotation to the
 SRSO dummy return thunk

On Mon, Jul 08, 2024 at 09:43:05AM -0700, Jim Mattson wrote:
> The linux-5.10-y backport of commit b377c66ae350 ("x86/retpoline: Add
> NOENDBR annotation to the SRSO dummy return thunk") misplaced the new
> NOENDBR annotation, repeating the annotation on __x86_return_thunk,
> rather than adding the annotation to the !CONFIG_CPU_SRSO version of
> srso_alias_untrain_ret, as intended.
> 
> Move the annotation to the right place.
> 
> Fixes: 0bdc64e9e716 ("x86/retpoline: Add NOENDBR annotation to the SRSO dummy return thunk")
> Reported-by: Greg Thelen <gthelen@...gle.com>
> Signed-off-by: Jim Mattson <jmattson@...gle.com>
> ---
>  arch/x86/lib/retpoline.S | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/lib/retpoline.S b/arch/x86/lib/retpoline.S
> index ab9b047790dd..d1902213a0d6 100644
> --- a/arch/x86/lib/retpoline.S
> +++ b/arch/x86/lib/retpoline.S
> @@ -105,6 +105,7 @@ __EXPORT_THUNK(srso_alias_untrain_ret)
>  /* dummy definition for alternatives */
>  SYM_START(srso_alias_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
>  	ANNOTATE_UNRET_SAFE
> +	ANNOTATE_NOENDBR
>  	ret
>  	int3
>  SYM_FUNC_END(srso_alias_untrain_ret)
> @@ -258,7 +259,6 @@ SYM_CODE_START(__x86_return_thunk)
>  	UNWIND_HINT_FUNC
>  	ANNOTATE_NOENDBR
>  	ANNOTATE_UNRET_SAFE
> -	ANNOTATE_NOENDBR

Whoops, those two here didn't look right, thanks for catching this.

Acked-by: Borislav Petkov (AMD) <bp@...en8.de>

And looking at 6.1, it is wrong there too. :-\

commit 7ef6a7f9b32fdfc8bec0a10e6d5ac5374d4f02e7
Author: Borislav Petkov (AMD) <bp@...en8.de>
Date:   Fri Apr 5 16:46:37 2024 +0200

    x86/retpoline: Add NOENDBR annotation to the SRSO dummy return thunk
    
    commit b377c66ae3509ccea596512d6afb4777711c4870 upstream.

Greg (KH - there's another Greg on CC too :-)), do you prefer such stable-only
fixes like that or should we revert the wrong backports and redo them?

Thx!

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette