Re: [PATCH 2/2] arm64: kexec_file: use more system keyrings to verify kernel image signature

[Will Deacon]

On Mon, Sep 27, 2021 at 08:50:04AM +0800, Coiby Xu wrote:
> From: Coiby Xu <coxu@xxxxxxxxxx>
> 
> This allows to verify arm64 kernel image signature using not only
> .builtin_trusted_keys but also .secondary_trusted_keys and .platform keyring.
> 
> Signed-off-by: Coiby Xu <coiby.xu@xxxxxxxxx>
> ---
>  arch/arm64/kernel/kexec_image.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/arch/arm64/kernel/kexec_image.c b/arch/arm64/kernel/kexec_image.c
> index 9ec34690e255..2357ee2f229a 100644
> --- a/arch/arm64/kernel/kexec_image.c
> +++ b/arch/arm64/kernel/kexec_image.c
> @@ -14,7 +14,6 @@
>  #include <linux/kexec.h>
>  #include <linux/pe.h>
>  #include <linux/string.h>
> -#include <linux/verification.h>
>  #include <asm/byteorder.h>
>  #include <asm/cpufeature.h>
>  #include <asm/image.h>
> @@ -133,8 +132,7 @@ static void *image_load(struct kimage *image,
>  #ifdef CONFIG_KEXEC_IMAGE_VERIFY_SIG
>  static int image_verify_sig(const char *kernel, unsigned long kernel_len)
>  {
> -	return verify_pefile_signature(kernel, kernel_len, NULL,
> -				       VERIFYING_KEXEC_PE_SIGNATURE);
> +	return arch_kexec_kernel_verify_pe_sig(kernel, kernel_len);

I'm fine with this in principle, but it looks like the first patch is the
important one.

So for the arm64 bit:

Acked-by: Will Deacon <will@xxxxxxxxxx>

Will