[v3,12/75] x86/boot/compressed/64: Switch to __KERNEL_CS after GDT is loaded
diff mbox series

Message ID 20200428151725.31091-13-joro@8bytes.org
State New
Headers show
Series
  • x86: SEV-ES Guest Support
Related show

Commit Message

Joerg Roedel April 28, 2020, 3:16 p.m. UTC
From: Joerg Roedel <jroedel@suse.de>

When the pre-decompression code loads its first GDT in startup_64, it is
still running on the CS value of the previous GDT. In the case of SEV-ES
this is the EFI GDT.

To make exception handling work (especially IRET) the CPU needs to
switch to a CS value in the current GDT, so jump to __KERNEL_CS after
the first GDT is loaded.

Signed-off-by: Joerg Roedel <jroedel@suse.de>
---
 arch/x86/boot/compressed/head_64.S | 8 ++++++++
 1 file changed, 8 insertions(+)

Comments

Borislav Petkov May 4, 2020, 10:41 a.m. UTC | #1
On Tue, Apr 28, 2020 at 05:16:22PM +0200, Joerg Roedel wrote:
> From: Joerg Roedel <jroedel@suse.de>
> 
> When the pre-decompression code loads its first GDT in startup_64, it is
> still running on the CS value of the previous GDT. In the case of SEV-ES
> this is the EFI GDT.
> 
> To make exception handling work (especially IRET) the CPU needs to
> switch to a CS value in the current GDT, so jump to __KERNEL_CS after
> the first GDT is loaded.
> 
> Signed-off-by: Joerg Roedel <jroedel@suse.de>
> ---
>  arch/x86/boot/compressed/head_64.S | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/arch/x86/boot/compressed/head_64.S b/arch/x86/boot/compressed/head_64.S
> index 4f7e6b84be07..6b11060c3a0f 100644
> --- a/arch/x86/boot/compressed/head_64.S
> +++ b/arch/x86/boot/compressed/head_64.S
> @@ -393,6 +393,14 @@ SYM_CODE_START(startup_64)
>  	addq	%rax, 2(%rax)
>  	lgdt	(%rax)
>  
> +	/* Reload CS so IRET returns to a CS actually in the GDT */
> +	pushq	$__KERNEL_CS
> +	leaq	.Lon_kernel_cs(%rip), %rax
> +	pushq	%rax
> +	lretq
> +
> +.Lon_kernel_cs:
> +
>  	/*
>  	 * paging_prepare() sets up the trampoline and checks if we need to
>  	 * enable 5-level paging.
> -- 

So I'm thinking I should take this one even now on the grounds that
it sanitizes CS to something known-good than what was there before and
who knows what set it and loaded the kernel...?

And that is a good thing in itself.
Joerg Roedel May 4, 2020, 11:27 a.m. UTC | #2
On Mon, May 04, 2020 at 12:41:29PM +0200, Borislav Petkov wrote:
> On Tue, Apr 28, 2020 at 05:16:22PM +0200, Joerg Roedel wrote:
> > +	/* Reload CS so IRET returns to a CS actually in the GDT */
> > +	pushq	$__KERNEL_CS
> > +	leaq	.Lon_kernel_cs(%rip), %rax
> > +	pushq	%rax
> > +	lretq
> > +
> > +.Lon_kernel_cs:
> > +
> >  	/*
> >  	 * paging_prepare() sets up the trampoline and checks if we need to
> >  	 * enable 5-level paging.
> > -- 
> 
> So I'm thinking I should take this one even now on the grounds that
> it sanitizes CS to something known-good than what was there before and
> who knows what set it and loaded the kernel...?
> 
> And that is a good thing in itself.

Right, sure. CS is basically undefined at this point and depends on what
loaded the kernel (EFI, legacy boot code, some container runtime...), so
setting it to something known is definitly good.

Regards,

	Joerg

Patch
diff mbox series

diff --git a/arch/x86/boot/compressed/head_64.S b/arch/x86/boot/compressed/head_64.S
index 4f7e6b84be07..6b11060c3a0f 100644
--- a/arch/x86/boot/compressed/head_64.S
+++ b/arch/x86/boot/compressed/head_64.S
@@ -393,6 +393,14 @@  SYM_CODE_START(startup_64)
 	addq	%rax, 2(%rax)
 	lgdt	(%rax)
 
+	/* Reload CS so IRET returns to a CS actually in the GDT */
+	pushq	$__KERNEL_CS
+	leaq	.Lon_kernel_cs(%rip), %rax
+	pushq	%rax
+	lretq
+
+.Lon_kernel_cs:
+
 	/*
 	 * paging_prepare() sets up the trampoline and checks if we need to
 	 * enable 5-level paging.