[PATCH] x86/kexec: Only write through identity mapping of control page

[Ning, Hongyu]

On 2024/12/13 4:11, David Woodhouse wrote:
> From: David Woodhouse <dwmw at amazon.co.uk>
> 
> The virtual mapping of the control page may have been _PAGE_GLOBAL and
> thus its PTE might not have been flushed on the %cr3 switch and it might
> effectively still be read-only. Move the writes to it down into the
> identity_mapped() function where the same %rip-relative addressing will
> get the new mapping.
> 
> The stack is fine, as that's using the identity mapped address anyway.
> 
> Fixes: 5a82223e0743 ("x86/kexec: Mark relocate_kernel page as ROX instead of RWX")
> Reported-by: Nathan Chancellor <nathan at kernel.org>
> Reported-by: "Ning, Hongyu" <hongyu.ning at linux.intel.com>
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219592
> Signed-off-by: David Woodhouse <dwmw at amazon.co.uk>

Tested-by: "Ning, Hongyu" <hongyu.ning at linux.intel.com>

> ---
>   arch/x86/kernel/relocate_kernel_64.S | 32 +++++++++++++++++-----------
>   1 file changed, 20 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
> index 553d67845b84..b9c80b3091c8 100644
> --- a/arch/x86/kernel/relocate_kernel_64.S
> +++ b/arch/x86/kernel/relocate_kernel_64.S
> @@ -90,22 +90,17 @@ SYM_CODE_START_NOALIGN(relocate_kernel)
>   	movq	kexec_pa_table_page(%rip), %r9
>   	movq	%r9, %cr3
>   
> -	/* Save %rsp and CRs. */
> -	movq    %rsp, saved_rsp(%rip)
> -	movq	%rax, saved_cr3(%rip)
> -	movq	%cr0, %rax
> -	movq	%rax, saved_cr0(%rip)
> -	/* Leave CR4 in %r13 to enable the right paging mode later. */
> -	movq	%cr4, %r13
> -	movq	%r13, saved_cr4(%rip)
> -
> -	/* save indirection list for jumping back */
> -	movq	%rdi, pa_backup_pages_map(%rip)
> +	/*
> +	 * The control page still might not be writable because the original
> +	 * kernel PTE may have had the _PAGE_GLOBAL bit set. Don't write to
> +	 * it except through the *identmap* address.
> +	 */
>   
>   	/* Save the preserve_context to %r11 as swap_pages clobbers %rcx. */
>   	movq	%rcx, %r11
>   
>   	/* setup a new stack at the end of the physical control page */
> +	movq	%rsp, %rbp
>   	lea	PAGE_SIZE(%rsi), %rsp
>   
>   	/* jump to identity mapped page */
> @@ -118,6 +113,19 @@ SYM_CODE_END(relocate_kernel)
>   
>   SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
>   	UNWIND_HINT_END_OF_STACK
> +
> +	/* Save original %rsp and CRs. */
> +	movq    %rbp, saved_rsp(%rip)
> +	movq	%rax, saved_cr3(%rip)
> +	movq	%cr0, %rax
> +	movq	%rax, saved_cr0(%rip)
> +	/* Leave CR4 in %r13 to enable the right paging mode later. */
> +	movq	%cr4, %r13
> +	movq	%r13, saved_cr4(%rip)
> +
> +	/* save indirection list for jumping back */
> +	movq	%rdi, pa_backup_pages_map(%rip)
> +
>   	/*
>   	 * %rdi	indirection page
>   	 * %rdx start address
> @@ -185,7 +193,7 @@ SYM_CODE_START_LOCAL_NOALIGN(identity_mapped)
>   	 *  - Machine check exception on TDX guest, if it was enabled before.
>   	 *    Clearing MCE might not be allowed in TDX guests, depending on setup.
>   	 *
> -	 * Use R13 that contains the original CR4 value, read in relocate_kernel().
> +	 * Use R13 that contains the original CR4 value
>   	 * PAE is always set in the original CR4.
>   	 */
>   	andl	$(X86_CR4_PAE | X86_CR4_LA57), %r13d