Missing SLS int3 in JMP_NOSPEC? (Was: [PATCH v2 6/6] x86: Add straight-line-speculation mitigation)

[Maciej S. Szmigiero]

On 4.12.2021 14:43, Peter Zijlstra wrote:
> Make use of an upcomming GCC feature to mitigate
> straight-line-speculation for x86:
>
>    https://gcc.gnu.org/g:53a643f8568067d7700a9f2facc8ba39974973d3
>    https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102952
>    https://bugs.llvm.org/show_bug.cgi?id=52323
>
> It's built tested on x86_64-allyesconfig using GCC-12 and GCC-11.
>
> Maintenace overhead of this should be fairly low due to objtool
> validation.
>
> Size overhead of all these additional int3 instructions comes to:
>
>     text	   data	    bss	    dec	    hex	filename
> 22267751	6933356	2011368	31212475	1dc43bb	defconfig-build/vmlinux
> 22804126	6933356	1470696	31208178	1dc32f2	defconfig-build/vmlinux.sls
>
> Or roughly 2.4% additional text.
>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
> ---
(..)
> --- a/arch/x86/lib/retpoline.S
> +++ b/arch/x86/lib/retpoline.S
> @@ -34,7 +34,7 @@ SYM_INNER_LABEL(__x86_indirect_thunk_\re
>   
>   	ALTERNATIVE_2 __stringify(ANNOTATE_RETPOLINE_SAFE; jmp *%\reg), \
>   		      __stringify(RETPOLINE \reg), X86_FEATURE_RETPOLINE, \
> -		      __stringify(lfence; ANNOTATE_RETPOLINE_SAFE; jmp *%\reg), X86_FEATURE_RETPOLINE_AMD
> +		      __stringify(lfence; ANNOTATE_RETPOLINE_SAFE; jmp *%\reg; int3), X86_FEATURE_RETPOLINE_AMD
>   
>   .endm
>   

Looking at this __x86_indirect_thunk_* change makes me wonder why there is
no similar int3 SLS protection in the X86_FEATURE_RETPOLINE_LFENCE case of
JMP_NOSPEC in arch/x86/include/asm/nospec-branch.h:
> .macro JMP_NOSPEC reg:req
> #ifdef CONFIG_RETPOLINE
> 	ALTERNATIVE_2 __stringify(ANNOTATE_RETPOLINE_SAFE; jmp *%\reg), \
> 		      __stringify(jmp __x86_indirect_thunk_\reg), X86_FEATURE_RETPOLINE, \
> 		      __stringify(lfence; ANNOTATE_RETPOLINE_SAFE; jmp *%\reg), X86_FEATURE_RETPOLINE_LFENCE
> #else

JMP_NOSPEC users seem to have no explicit trailing int3 instructions
either.

Or am I missing something here?

Thanks,
Maciej