Re: [PATCH v7 02/12] x86: Use symbol name in jump table for PIE support

[Thomas Garnier]

On Mon, May 20, 2019 at 4:20 PM Thomas Garnier <thgarnie@xxxxxxxxxxxx> wrote:
>
> From: Thomas Garnier <thgarnie@xxxxxxxxxx>
>
> Replace the %c constraint with %P. The %c is incompatible with PIE
> because it implies an immediate value whereas %P reference a symbol.
> Change the _ASM_PTR reference to .long for expected relocation size and
> add a long padding to ensure entry alignment.
>
> Position Independent Executable (PIE) support will allow to extend the
> KASLR randomization range below 0xffffffff80000000.
>
> Signed-off-by: Thomas Garnier <thgarnie@xxxxxxxxxx>
> ---
>  arch/x86/include/asm/jump_label.h | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/arch/x86/include/asm/jump_label.h b/arch/x86/include/asm/jump_label.h
> index 65191ce8e1cf..e47fad8ee632 100644
> --- a/arch/x86/include/asm/jump_label.h
> +++ b/arch/x86/include/asm/jump_label.h
> @@ -25,9 +25,9 @@ static __always_inline bool arch_static_branch(struct static_key *key, bool bran
>                 ".pushsection __jump_table,  \"aw\" \n\t"
>                 _ASM_ALIGN "\n\t"
>                 ".long 1b - ., %l[l_yes] - . \n\t"
> -               _ASM_PTR "%c0 + %c1 - .\n\t"
> +               _ASM_PTR "%P0 - .\n\t"
>                 ".popsection \n\t"
> -               : :  "i" (key), "i" (branch) : : l_yes);
> +               : :  "X" (&((char *)key)[branch]) : : l_yes);
>
>         return false;
>  l_yes:
> @@ -42,9 +42,9 @@ static __always_inline bool arch_static_branch_jump(struct static_key *key, bool
>                 ".pushsection __jump_table,  \"aw\" \n\t"
>                 _ASM_ALIGN "\n\t"
>                 ".long 1b - ., %l[l_yes] - . \n\t"
> -               _ASM_PTR "%c0 + %c1 - .\n\t"
> +               _ASM_PTR "%P0 - .\n\t"
>                 ".popsection \n\t"
> -               : :  "i" (key), "i" (branch) : : l_yes);
> +               : : "X" (&((char *)key)[branch]) : : l_yes);
>
>         return false;
>  l_yes:
> --
> 2.21.0.1020.gf2820cf01a-goog
>

Realized I forgot to address a feedback from the previous iteration on
this specific patch. Ignore it I will work to check if it can be
remove on the next iteration.


-- 
Thomas