Re: [RFC PATCH] x86/head_64.S: remove redundant check that kernel address is 2M aligned

[Brian Gerst]

On Fri, Jan 22, 2016 at 1:13 PM, Alexander Kuleshov
<kuleshovmail@xxxxxxxxx> wrote:
> We check that the base address of the kernel is 2M aligned in
> the arch/x86/kernel/head_65.S right after jump to the decompressed
> kernel. But we already have a check in the decompress_kernel()
> which validates that kernel location is MIN_KERNEL_ALIGN aligned
> which is 2M too for x86_64.
>
> Signed-off-by: Alexander Kuleshov <kuleshovmail@xxxxxxxxx>
> ---
>  arch/x86/kernel/head_64.S | 6 ------
>  1 file changed, 6 deletions(-)
>
> diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
> index ffdc0e8..4967cba 100644
> --- a/arch/x86/kernel/head_64.S
> +++ b/arch/x86/kernel/head_64.S
> @@ -75,12 +75,6 @@ startup_64:
>         leaq    _text(%rip), %rbp
>         subq    $_text - __START_KERNEL_map, %rbp
>
> -       /* Is the address not 2M aligned? */
> -       movq    %rbp, %rax
> -       andl    $~PMD_PAGE_MASK, %eax
> -       testl   %eax, %eax
> -       jnz     bad_address
> -
>         /*
>          * Is the address too large?
>          */

I think we still need to do the check, in case we came from a 64-bit
bootloader that directly jumped to startup_64.  However, this check
can be simplified to:

    testl $~PMD_PAGE_MASK, %ebp
    jnz bad_address

--
Brian Gerst